CN115346263A - Eye track anti-fraud method, device, equipment and medium based on artificial intelligence - Google Patents
Eye track anti-fraud method, device, equipment and medium based on artificial intelligence Download PDFInfo
- Publication number
- CN115346263A CN115346263A CN202210987439.2A CN202210987439A CN115346263A CN 115346263 A CN115346263 A CN 115346263A CN 202210987439 A CN202210987439 A CN 202210987439A CN 115346263 A CN115346263 A CN 115346263A
- Authority
- CN
- China
- Prior art keywords
- track
- gaze
- positive
- negative
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
- G06Q30/0185—Product, service or business identity fraud
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/197—Matching; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Ophthalmology & Optometry (AREA)
- Human Computer Interaction (AREA)
- Databases & Information Systems (AREA)
- Accounting & Taxation (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Entrepreneurship & Innovation (AREA)
- Computing Systems (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Finance (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Image Analysis (AREA)
Abstract
The application provides a gaze track anti-fraud method and device based on artificial intelligence, an electronic device and a storage medium, wherein the gaze track anti-fraud method based on artificial intelligence comprises the following steps: acquiring eye video data to generate an eye sample track set; segmenting the gaze sample track set to obtain a track initial segment set; screening the initial track fragment set to obtain a positive track effective fragment set and a negative track effective fragment set; constructing a gaze track fraud recognizer based on the positive track valid segment set and the negative track valid segment set; performing anti-fraud detection on a target gaze track to be detected based on the gaze track fraud identifier to obtain a detection result; and updating the gaze track fraud identifier based on the detection result to improve the detection accuracy of the gaze track fraud identifier. By generating the eye track fraud identifier, the accuracy of identifying the eye track fraud risk is improved.
Description
Technical Field
The present application relates to the field of artificial intelligence technologies, and in particular, to a gaze track anti-fraud method and apparatus, an electronic device, and a storage medium based on artificial intelligence.
Background
Fraud risk identification is a wind control direction which exists in the financial field for a long time and has high difficulty, in the traditional fraud risk identification, correlation analysis such as characters, browsing records, equipment IP and the like of a client is started, however, the methods have certain time-delay property and are easy to break, and the accuracy of fraud risk identification is low.
In various research and practical applications against fraud, there is very little risk of fraud through visual identification, and in related applications, micro-expression is an attempt based on video images, but the effect has been unstable and not accurate enough. Therefore, it becomes important how to effectively identify the fraud risk present in the gaze track by means of visual techniques.
Disclosure of Invention
In view of the above, there is a need to provide a gaze track anti-fraud method, apparatus, electronic device and storage medium based on artificial intelligence, so as to solve the technical problem of how to improve the accuracy of fraud risk identification in the financial loan field.
The application provides a gaze track anti-fraud method based on artificial intelligence, which comprises the following steps:
acquiring eye video data to generate an eye sample trajectory set, wherein the eye sample trajectory set comprises an eye positive sample trajectory set and an eye negative sample trajectory set;
all gaze sample tracks in the gaze sample track set are segmented to obtain a track initial segment set, wherein the track initial segment set comprises a positive track initial segment set and a negative track initial segment set;
screening the initial track fragment set to obtain a positive track effective fragment set and a negative track effective fragment set;
constructing a gaze track fraud recognizer based on the positive track valid segment set and the negative track valid segment set;
performing anti-fraud detection on a target gaze track to be detected based on the gaze track fraud identifier to obtain a detection result;
in some embodiments, said obtaining video data to generate a gaze sample trajectory set comprises:
collecting eye sight video data;
carrying out positive and negative sample marking on all gaze video data according to a preset mode, taking all gaze video data with positive sample marks as a gaze positive sample data set, and taking all gaze video data with negative sample marks as a gaze negative sample data set;
generating eye gaze positive sample tracks based on the eye gaze video data in the eye gaze positive sample data set, and taking all eye gaze positive sample tracks as an eye gaze positive sample track set;
generating gaze negative sample tracks based on gaze video data in the gaze negative sample data set, and taking all gaze negative sample tracks as a gaze negative sample track set;
and taking the gaze positive sample trajectory set and the gaze negative sample trajectory set as gaze sample trajectory sets.
In some embodiments, the segmenting all gaze sample trajectories in the gaze sample trajectory set to obtain an initial set of trajectory segments comprises:
extracting low-frequency information of each gaze sample track in the gaze sample track set to obtain a gaze sample low-frequency track;
calculating extreme points of the low-frequency track of the gaze sample, and segmenting the low-frequency track of the gaze sample according to all the extreme points to obtain a plurality of track initial segments of the low-frequency track of the gaze sample;
respectively taking all track initial segments corresponding to the gaze sample tracks in the gaze positive sample track set and the gaze negative sample track set as a positive track initial segment set and a negative track initial segment set;
and taking the positive track initial segment set and the negative track initial segment set as track initial segment sets.
In some embodiments, the screening the initial set of track segments to obtain a positive set of track valid segments and a negative set of track valid segments includes:
calculating the segment similarity of a target positive track initial segment and each positive track initial segment in the positive track initial segment set, wherein the target positive track initial segment is any one of the positive track initial segment set;
obtaining a positive track similar segment set corresponding to the target positive track initial segment based on a preset threshold and the segment similarity;
traversing each positive track initial segment in the positive track initial segment set to obtain a positive track similar segment set corresponding to each positive track initial segment;
calculating the segment similarity of a target negative track initial segment and each negative track initial segment in the negative track initial segment set, wherein the target negative track initial segment is any one of the negative track initial segment set;
obtaining a negative track similar segment set corresponding to the target negative track initial segment based on a preset threshold and the segment similarity;
traversing each negative track initial segment in the negative track initial segment set to obtain a negative track similar segment set corresponding to each negative track initial segment;
and respectively screening all the positive track similar segment sets and all the negative track similar segment sets to obtain a positive track effective segment set and a negative track effective segment set.
In some embodiments, the screening all positive trace-like segment sets and all negative trace-like segment sets respectively to obtain positive trace-valid segment sets and negative trace-valid segment sets includes:
respectively counting the number of the segments included in each positive track similar segment set and each negative track similar segment set;
sequencing all positive track similar segment sets and all negative track similar segment sets respectively according to the sequence of the segment numbers from large to small to obtain positive track sequencing results and negative track sequencing results;
and respectively screening all the positive track similar segment sets and all the negative track similar segment sets based on the positive track sorting result and the negative track sorting result to obtain a plurality of positive track effective segment sets and negative track effective segment sets with the same quantity.
In some embodiments, said constructing a gaze track fraud identifier based on said set of positive track valid segments and said set of negative track valid segments comprises:
extracting the longest common sequence of all the segments in each positive track effective segment set;
weighting and summing the longest common sequence of all positive track effective fragment sets to obtain a positive track reference template;
extracting the longest common sequence of all the segments in each negative track effective segment set;
carrying out weighted summation on the longest common sequence of all negative track effective fragment sets to obtain a negative track reference template;
and taking the positive track reference template and the negative track reference template as eye track fraud identifiers.
In some embodiments, the detecting, based on the gaze track fraud identifier, performing fraud prevention detection on the target gaze track to be detected to obtain a detection result includes:
respectively calculating the similarity between a target gaze track to be detected and a positive track reference template and a negative track reference template in the gaze track fraud identifier to obtain a positive track similarity and a negative track similarity;
if the positive track similarity is greater than the negative track similarity, the target gaze track to be detected has a fraud risk, and the detection result is a fraud high risk;
and if the positive track similarity is not greater than the negative track similarity, the target gaze track to be detected does not have fraud risk, and the detection result is fraud low risk.
The embodiment of the present application further provides an anti-cheating device of sight track based on artificial intelligence, the device includes:
an obtaining unit configured to obtain gaze video data to generate a gaze sample trajectory set, the gaze sample trajectory set including a gaze positive sample trajectory set and a gaze negative sample trajectory set;
the segmentation unit is used for segmenting all gaze sample tracks in the gaze sample track set to obtain a track initial segment set, wherein the track initial segment set comprises a positive track initial segment set and a negative track initial segment set;
the screening unit is used for screening the initial track segment set to obtain a positive track effective segment set and a negative track effective segment set;
the construction unit is used for constructing a gaze track fraud recognizer based on the positive track effective segment set and the negative track effective segment set;
and the detection unit is used for carrying out anti-cheating detection on the target gaze track to be detected based on the gaze track cheating identifier to obtain a detection result.
An embodiment of the present application further provides an electronic device, where the electronic device includes:
a memory storing at least one instruction;
and the processor executes the instructions stored in the memory to realize the artificial intelligence based gaze track anti-fraud method.
The embodiment of the application also provides a computer-readable storage medium, and at least one instruction is stored in the computer-readable storage medium and executed by a processor in the electronic device to implement the artificial intelligence based gaze track anti-fraud method.
According to the method and the device, the positive and negative eye sample tracks are extracted from the obtained eye video data and are combined after the positive and negative eye sample tracks are segmented, so that the eye track fraud identifier capable of detecting whether the eye track contains fraud information is obtained, and the accuracy of identifying the fraud risk of the time sequence eye track is effectively improved.
Drawings
FIG. 1 is a flow chart of a preferred embodiment of an artificial intelligence based gaze track anti-fraud method to which the present application relates.
Fig. 2 is a functional block diagram of a preferred embodiment of an artificial intelligence based gaze track anti-fraud device to which the present application relates.
Fig. 3 is a schematic structural diagram of an electronic device according to a preferred embodiment of the artificial intelligence-based gaze track anti-fraud method.
Fig. 4 is a schematic diagram of a gaze sample trajectory to which the present application relates.
FIG. 5 is a schematic diagram of the calculation of the dynamic time warping algorithm to which the present application relates.
Detailed Description
For a clearer understanding of the objects, features and advantages of the present application, reference is made to the following detailed description of the present application along with the accompanying drawings and specific examples. It should be noted that the embodiments and features of the embodiments of the present application may be combined with each other without conflict. In the following description, numerous specific details are set forth to provide a thorough understanding of the present application, and the described embodiments are merely a subset of the embodiments of the present application and are not intended to be a complete embodiment.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", may explicitly or implicitly include one or more of the described features. In the description of the present application, "a plurality" means two or more unless specifically limited otherwise.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the description of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
The embodiment of the Application provides an artificial intelligence-based gaze track anti-fraud method, which can be applied to one or more electronic devices, wherein the electronic devices are devices capable of automatically performing numerical calculation and/or information processing according to preset or stored instructions, and hardware of the electronic devices includes but is not limited to a microprocessor, an Application Specific Integrated Circuit (ASIC), a Programmable Gate Array (FPGA), a Digital Signal Processor (DSP), an embedded device, and the like.
The electronic device may be any electronic product capable of performing human-computer interaction with a client, for example, a Personal computer, a tablet computer, a smart phone, a Personal Digital Assistant (PDA), a game machine, an Internet Protocol Television (IPTV), an intelligent wearable device, and the like.
The electronic device may also include a network device and/or a client device. The network device includes, but is not limited to, a single network server, a server group consisting of a plurality of network servers, or a Cloud Computing (Cloud Computing) based Cloud consisting of a large number of hosts or network servers.
The Network where the electronic device is located includes, but is not limited to, the internet, a wide area Network, a metropolitan area Network, a local area Network, a Virtual Private Network (VPN), and the like.
Fig. 1 is a flowchart of a preferred embodiment of the artificial intelligence-based gaze track anti-fraud method according to the present application. The order of the steps in the flow chart may be changed and some steps may be omitted according to different needs.
S10, obtaining the eye video data to generate an eye sample track set, wherein the eye sample track set comprises an eye positive sample track set and an eye negative sample track set.
In an alternative embodiment, the obtaining the gaze video data to generate the gaze sample track set comprises:
s101, collecting eye sight video data;
s102, marking positive and negative samples of all gaze video data according to a preset mode, taking all gaze video data with positive sample marks as a gaze positive sample data set, and taking all gaze video data with negative sample marks as a gaze negative sample data set;
s103, generating gaze correction sample tracks based on gaze video data in the gaze correction sample data set, and taking all gaze correction sample tracks as a gaze correction sample track set;
s104, generating gaze negative sample tracks based on gaze video data in the gaze negative sample data set, and taking all gaze negative sample tracks as gaze negative sample track sets;
and S105, taking the gaze positive sample track set and the gaze negative sample track set as gaze sample track sets.
In the optional embodiment, in a large-amount transfer scene, a loan application scene and other scenes in the financial field, a user needs to pay attention to whether fraud risks exist in time, so that when the user transacts financial services remotely, such as applying the loan, the user needs to record video data in real time for auditing and filing, and massive recorded videos with the user sight data are collected as the sight video data in the scheme.
In this alternative embodiment, all the gaze video data may be manually marked with positive and negative samples according to whether fraud has occurred, so that all the gaze video data may be classified into two categories. All the gaze video data with the fraudulent conduct is marked as positive samples, all the gaze video data without the fraudulent conduct is marked as negative samples, all the gaze video data with the positive sample marks are used as a gaze positive sample data set, and all the gaze video data with the negative sample marks are used as a gaze negative sample data set.
In this optional embodiment, for the obtained gaze positive sample data set and gaze negative sample data set, a preset number of frames may be extracted from all gaze video data first, for example, one frame is extracted from every five consecutive frames as a target frame, so that the problem of too high system resource occupancy rate caused by processing a large number of video frames at the same time may be avoided.
In this optional embodiment, after a preset number of consecutive target frames are extracted from each gaze video data, the gaze focus in each target frame may be positioned by the microwave displacement sensor as the coordinates of the target point, and then the coordinates of the target points of all target frames are connected in sequence according to the time sequence, thereby generating the gaze movement trajectory. Generating gaze negative sample tracks based on gaze video data in the gaze negative sample data set, and taking all gaze negative sample tracks as gaze negative sample track sets, wherein the gaze negative sample tracks correspond to gaze video data in the gaze negative sample data set one by one; and meanwhile, generating a gaze correction sample track based on the gaze video data in the gaze correction sample data set, and taking all the gaze correction sample tracks as a gaze correction sample track set, wherein the gaze correction sample tracks correspond to the gaze video data in the gaze correction sample data set one by one.
In this alternative embodiment, the obtained gaze positive sample trajectory set and the obtained gaze negative sample trajectory set are used as gaze sample trajectory sets.
Therefore, positive and negative sight sample track sets with fraud labels can be obtained, and accurate data support is provided for a sight track fraud recognizer capable of detecting fraud information in a subsequent process.
S11, all gaze sample tracks in the gaze sample track set are segmented to obtain a track initial segment set, wherein the track initial segment set comprises a positive track initial segment set and a negative track initial segment set.
In an optional embodiment, the segmenting all gaze sample trajectories in the gaze sample trajectory set to obtain trajectory initial segment sets, the trajectory initial segment sets including a positive trajectory initial segment set and a negative trajectory initial segment set, includes:
s111, extracting low-frequency information of each gaze sample track in the gaze sample track set to obtain a gaze sample low-frequency track;
s112, calculating extreme points of the low-frequency track of the gaze sample, and segmenting the low-frequency track of the gaze sample according to all the extreme points to obtain a plurality of track initial segments of the low-frequency track of the gaze sample;
s113, respectively taking all track initial segments corresponding to the gaze sample tracks in the gaze positive sample track set and the gaze negative sample track set as a positive track initial segment set and a negative track initial segment set;
and S114, taking the positive track initial segment set and the negative track initial segment set as track initial segment sets.
In this optional embodiment, since the gaze track may change rapidly with time, and the carried high-frequency information is complex and variable, in order to effectively extract gaze track information having practical significance, low-frequency information of each gaze sample track in the gaze sample track set may be extracted through wavelet transformation. The wavelet transformation is a common signal frequency analysis tool, and can decompose each gaze sample track in the gaze sample track set to obtain high-frequency information and low-frequency information of each gaze sample track in the gaze sample track set. The reason for selecting the low-frequency information is that the high-frequency information carried by the gaze track changes frequently, so that effective gaze track characteristics cannot be extracted, and the low-frequency information changes less, so that the gaze track characteristics can be effectively reflected.
In this optional embodiment, for each obtained low-frequency trajectory of the gaze sample, the initial trajectory segments of the plurality of low-frequency trajectories of the gaze sample can be obtained by calculating the extreme point of each low-frequency trajectory of the gaze sample and segmenting the low-frequency trajectory of the gaze sample into a plurality of segments according to the extreme point. As shown in fig. 4, the obtained gaze sample trajectory is shown, where a black point A, B, C, D is an extreme point of a low-frequency trajectory of the gaze sample obtained after wavelet transform processing is performed on the gaze sample trajectory, and the gaze sample trajectory is finally divided into 5 initial trajectory segments.
In this optional embodiment, after all the gaze sample trajectories in the gaze positive sample trajectory set and the gaze negative sample trajectory set are segmented, all trajectory initial segments corresponding to the obtained gaze sample trajectories in the gaze positive sample trajectory set and the gaze negative sample trajectory set are used as a positive trajectory initial segment set and a negative trajectory initial segment set in the present solution, and the positive trajectory initial segment set and the negative trajectory initial segment set are used as trajectory initial segment sets.
Therefore, all the gaze sample tracks can be segmented, a large number of track initial segments with practical significance are obtained, and data support is provided for screening more effective track initial segments in the subsequent process.
And S12, screening the initial track segment set to obtain a positive track effective segment set and a negative track effective segment set.
In an optional embodiment, the screening the initial track segment set to obtain a positive track valid segment set and a negative track valid segment set includes:
s121, calculating the segment similarity of a target positive track initial segment and each positive track initial segment in the positive track initial segment set, wherein the target positive track initial segment is any one of the positive track initial segment set;
s122, obtaining a positive track similar segment set corresponding to the target positive track initial segment based on a preset threshold and the segment similarity;
s123, traversing each positive track initial segment in the positive track initial segment set to obtain a positive track similar segment set corresponding to each positive track initial segment;
s124, calculating the segment similarity of a target negative track initial segment and each negative track initial segment in the negative track initial segment set, wherein the target negative track initial segment is any one of the negative track initial segment set;
s125, obtaining a negative track similar segment set corresponding to the target negative track initial segment based on a preset threshold and the segment similarity;
s126, traversing each negative track initial segment in the negative track initial segment set to obtain a negative track similar segment set corresponding to each negative track initial segment;
and S127, screening all the positive track similar segment sets and all the negative track similar segment sets respectively to obtain positive track effective segment sets and negative track effective segment sets.
In this optional embodiment, the segment similarity between the target positive track initial segment and each positive track initial segment in the positive track initial segment set may be calculated according to a Dynamic Time Warping (DTW) algorithm, where the target positive track initial segment is any one of the positive track initial segment sets.
In this alternative embodiment, the DTW algorithm is a method for measuring similarity between two time sequences, and since different time sequences may only have a displacement in the time axis, the DTW algorithm can calculate the similarity between the two time sequences by extending and shortening the time sequences, compared with the conventional method for calculating the euclidean distance between the two time sequences. As shown in fig. 5, an upper solid line and a lower solid line respectively represent a target positive track initial segment and any one positive track initial segment in the positive track initial segment set except the target positive track initial segment, a dashed line between the two positive track initial segments represents similar points between the two positive track initial segments, and the DTW algorithm measures the similarity between the two positive track initial segments as a segment similarity by calculating a sum of distances between all the similar points. Wherein, because the DTW algorithm has a plurality of variants, the specific calculation process in the scheme can be consistent with the calculation process of the conventional DTW algorithm.
In this optional embodiment, after obtaining the segment similarity between the target positive track initial segment and each of the other positive track initial segments, filtering all the segment similarities corresponding to the target positive track initial segment based on a preset threshold, and taking the positive track initial segment corresponding to the segment similarity greater than the preset threshold as the positive track similar segment corresponding to the target positive track initial segment; and filtering the positive track initial segment corresponding to the segment similarity not greater than the preset threshold, and finally taking all positive track similar segments corresponding to the target positive track initial segment as a positive track similar segment set. The preset threshold of the segment similarity may be 0.6, that is, the positive track initial segment corresponding to the segment similarity greater than 0.6 is used as the positive track similar segment corresponding to the target positive track initial segment; and filtering out the positive track initial segment corresponding to the segment similarity not greater than 0.6.
In this optional embodiment, a positive track similar segment set corresponding to each positive track initial segment may be obtained by traversing each positive track initial segment in the positive track initial segment set. In accordance with the process of acquiring the positive track similar segment set, a negative track similar segment set corresponding to each negative track initial segment can be calculated and acquired.
In this alternative embodiment, all positive trajectory similar segment sets and all negative trajectory similar segment sets may be sorted according to the sequence of the number of segments from large to small, so as to obtain a positive trajectory sorting result and a negative trajectory sorting result. Then, a positive track similar segment set and a negative track similar segment set ranked in a previous preset order are respectively selected from the positive track sorting result and the negative track sorting result, for example, the previous preset order may be 50, that is, 50 positive track similar segment sets and 50 negative track similar segment sets are finally obtained according to the number of the segments. In the scheme, each positive track similar segment set obtained through screening is used as a positive track effective segment set, and each negative track similar segment set is used as a negative track effective segment set.
Therefore, the initial track segment set can be screened according to the number of the segments, more effective initial track segments can be obtained, and the detection accuracy of the gaze track fraud recognizer generated in the subsequent process can be improved.
And S13, constructing a gaze track fraud recognizer based on the positive track effective segment set and the negative track effective segment set.
In an optional embodiment, said constructing a gaze track fraud identifier based on said set of positive track valid segments and said set of negative track valid segments comprises:
s131, extracting the longest common sequence of all the segments in each positive track effective segment set;
s132, carrying out weighted summation on the longest public sequence of all positive track effective segment sets to obtain a positive track reference template;
s133, extracting the longest common sequence of all the segments in each negative track effective segment set;
s134, carrying out weighted summation on the longest common sequences of all negative track effective segment sets to obtain a negative track reference template;
and S135, taking the positive track reference template and the negative track reference template as gaze track fraud identifiers.
In this optional embodiment, since a plurality of track valid segments similar to each other exist in each positive track valid segment set and negative track valid segment set, and the track valid segments are not completely the same, in order to obtain the common characteristics of all track valid segments in each positive track valid segment set and negative track valid segment set, the longest common sequence of all track valid segments in each positive track valid segment set and negative track valid segment set is extracted by a Longest Common Subsequence (LCS) algorithm in the present solution.
In this alternative embodiment, the lcs algorithm is used to find the longest subsequence of all sequences in a sequence set, such as str1= "ABCBDCA", str2= "DABDA", one common subsequence of both being AA and the longest common subsequence being ABDA. Wherein the characters in the common subsequence need not be contiguous in the original sequence.
In this optional embodiment, after the longest common subsequence of all the segments in each positive track effective segment set and each negative track effective segment set is extracted, weights may be given to the corresponding longest common subsequence according to the number of track effective segments included in the positive track effective segment set or the negative track effective segment set in which each longest common subsequence is located, and the longest common subsequences of all the positive track effective segment sets and all the negative track effective segment sets for which weights are obtained are weighted and summed, so that a positive track reference template and a negative track reference template are obtained respectively.
In this optional embodiment, taking the acquisition of the positive track reference templates as an example, a specific process of giving a weight to the longest common subsequence corresponding to each positive track reference template is as follows: the total number of positive track effective fragments included in a positive track effective fragment set in which each longest public subsequence is located is counted, all the total numbers are summed to obtain the total number of fragments, and the ratio of the total number of positive track effective fragments corresponding to each longest public subsequence to the total number of fragments is used as the weight of the corresponding longest public subsequence.
In this optional embodiment, the process of obtaining the negative track reference template in the present embodiment is the same as the process of obtaining the positive track reference template, and the finally obtained positive track reference template and the negative track reference template are used together as a gaze track fraud identifier for performing anti-fraud detection on a gaze track to be detected.
Therefore, the gaze track fraud recognizer comprising the positive track reference template and the negative track reference template can be generated, and the fraud risk of the gaze track can be accurately recognized.
And S14, performing anti-fraud detection on the target gaze track to be detected based on the gaze track fraud identifier to obtain a detection result.
In an optional embodiment, the performing, based on the gaze track fraud identifier, fraud detection on the target gaze track to be detected to obtain a detection result includes:
s141, respectively calculating the similarity between the target gaze track to be detected and the positive track reference template and the negative track reference template in the gaze track fraud identifier to obtain a positive track similarity and a negative track similarity;
s142, if the positive track similarity is greater than the negative track similarity, the target gaze track to be detected has a fraud risk, and the detection result is a fraud high risk;
and S143, if the positive track similarity is not greater than the negative track similarity, the target gaze track to be detected does not have a fraud risk, and the detection result is a fraud low risk.
In this optional embodiment, for the target gaze track to be detected, the similarity between the target gaze track to be detected and the positive track reference template and the negative track reference template in the gaze track fraud identifier may be calculated respectively as the positive track similarity and the negative track similarity by using the DTW algorithm.
In this optional embodiment, the obtained positive track similarity and the negative track similarity are compared, and if the positive track similarity is greater than the negative track similarity, it is indicated that the target gaze track to be detected is closer to the positive track reference template in the gaze track fraud identifier, and a fraud risk is present, and the detection result is a fraud high risk; and if the positive track similarity is not greater than the negative track similarity, the target gaze track to be detected is closer to the negative track reference template in the gaze track fraud identifier, and the target gaze track to be detected does not have fraud risk, and the detection result is low risk of fraud.
Therefore, the target gaze track to be detected can be quickly detected and an accurate detection result can be obtained through the gaze track fraud identifier, and fraud risks of the target gaze track to be detected can be effectively distinguished.
In an alternative embodiment, the gaze track fraud identifier may be updated based on the detection result to improve the detection accuracy of the gaze track fraud identifier.
In this optional embodiment, the gaze track fraud identifier may be updated based on the detection result, and if the detection result is a high risk of fraud, the target gaze track corresponding to the high risk of fraud is marked as a gaze positive sample and added to the gaze positive sample track set, so that a new positive track reference template is generated through a series of processing procedures shown in this scheme to update the gaze track fraud identifier; and if the detection result is fraud low risk, marking the target gaze track corresponding to the fraud high risk as a gaze negative sample, adding the gaze negative sample track set into the gaze negative sample track set, and finally generating a new negative track reference template to update the gaze track fraud identifier.
In this way, the detection accuracy of the gaze track fraud identifier can be continuously improved by the continuous updating of the gaze track fraud identifier.
Referring to fig. 2, fig. 2 is a functional block diagram of a preferred embodiment of the artificial intelligence based gaze track anti-fraud apparatus of the present application. The artificial intelligence-based gaze track anti-fraud device 11 comprises an obtaining unit 110, a segmentation unit 111, a screening unit 112, a construction unit 113 and a detection unit 114. A module/unit as referred to herein is a series of computer readable instruction segments capable of being executed by the processor 13 and performing a fixed function, and is stored in the memory 12. In the present embodiment, the functions of the modules/units will be described in detail in the following embodiments.
In an alternative embodiment, the acquisition unit 110 is configured to acquire gaze video data to generate a gaze sample trajectory set comprising a gaze positive sample trajectory set and a gaze negative sample trajectory set.
In an alternative embodiment, the obtaining the gaze video data to generate the gaze sample track set comprises:
collecting eye sight video data;
carrying out positive and negative sample marking on all gaze video data according to a preset mode, taking all gaze video data with positive sample marks as a gaze positive sample data set, and taking all gaze video data with negative sample marks as a gaze negative sample data set;
generating eye positive sample tracks based on the eye video data in the eye positive sample data set, and taking all the eye positive sample tracks as an eye positive sample track set;
generating gaze negative sample tracks based on gaze video data in the gaze negative sample data set, and taking all gaze negative sample tracks as a gaze negative sample track set;
and taking the gaze positive sample trajectory set and the gaze negative sample trajectory set as gaze sample trajectory sets.
In the optional embodiment, in a large-amount transfer scene, a loan application scene and other scenes in the financial field, a user needs to pay attention to whether a fraud risk exists or not in time, so that when the user transacts financial services remotely, such as applying for loan, the user needs to record video data in real time for auditing and filing, and in the scheme, massive recorded video with the client sight data is collected to serve as the sight video data.
In this alternative embodiment, all the gaze video data may be manually marked with positive and negative samples according to whether fraud has occurred, so that all the gaze video data may be classified into two categories. All the gaze video data with the fraudulent conduct is marked as positive samples, all the gaze video data without the fraudulent conduct is marked as negative samples, all the gaze video data with the positive sample marks are used as a gaze positive sample data set, and all the gaze video data with the negative sample marks are used as a gaze negative sample data set.
In this optional embodiment, for the obtained gaze positive sample data set and gaze negative sample data set, a preset number of frames may be extracted from all gaze video data, for example, one frame is extracted from every five consecutive frames as a target frame, so that the problem of too high system resource occupancy rate caused by processing a large number of video frames at the same time may be avoided.
In this optional embodiment, after a preset number of consecutive target frames are extracted from each gaze video data, the gaze focus in each target frame may be positioned by the microwave displacement sensor as the coordinates of the target point, and then the coordinates of the target points of all target frames are connected in sequence according to the time sequence, thereby generating the gaze movement trajectory. Generating gaze negative sample tracks based on gaze video data in the gaze negative sample data set, and taking all gaze negative sample tracks as gaze negative sample track sets, wherein the gaze negative sample tracks correspond to gaze video data in the gaze negative sample data set one by one; and meanwhile, generating a gaze correction sample track based on the gaze video data in the gaze correction sample data set, and taking all the gaze correction sample tracks as a gaze correction sample track set, wherein the gaze correction sample tracks correspond to the gaze video data in the gaze correction sample data set one by one.
In this alternative embodiment, the obtained gaze positive sample trajectory set and gaze negative sample trajectory set are used as gaze sample trajectory sets.
In an optional embodiment, the slicing unit 111 is configured to slice all gaze sample trajectories in the gaze sample trajectory set to obtain an initial trajectory segment set, where the initial trajectory segment set includes a positive initial trajectory segment set and a negative initial trajectory segment set.
In an optional embodiment, the segmenting all gaze sample trajectories in the gaze sample trajectory set to obtain trajectory initial segment sets, the trajectory initial segment sets including a positive trajectory initial segment set and a negative trajectory initial segment set, includes:
extracting low-frequency information of each gaze sample track in the gaze sample track set to obtain a gaze sample low-frequency track;
calculating extreme points of the low-frequency track of the gaze sample, and segmenting the low-frequency track of the gaze sample according to all the extreme points to obtain a plurality of track initial segments of the low-frequency track of the gaze sample;
respectively taking all track initial segments corresponding to the gaze sample tracks in the gaze positive sample track set and the gaze negative sample track set as a positive track initial segment set and a negative track initial segment set;
and taking the positive track initial segment set and the negative track initial segment set as track initial segment sets.
In this optional embodiment, since the gaze track may change rapidly with time, and the carried high-frequency information is complex and variable, in order to effectively extract gaze track information having practical significance, low-frequency information of each gaze sample track in the gaze sample track set may be extracted through wavelet transformation. The wavelet transformation is a common signal frequency analysis tool, and can decompose each gaze sample track in the gaze sample track set to obtain high-frequency information and low-frequency information of each gaze sample track in the gaze sample track set. The reason for selecting the low-frequency information is that the high-frequency information carried by the gaze track changes frequently, so that effective gaze track characteristics cannot be extracted, and the low-frequency information changes less, so that the gaze track characteristics can be effectively reflected.
In this optional embodiment, for each obtained low-frequency trajectory of the gaze sample, the initial trajectory segments of the plurality of low-frequency trajectories of the gaze sample can be obtained by calculating the extreme point of each low-frequency trajectory of the gaze sample and segmenting the low-frequency trajectory of the gaze sample into a plurality of segments according to the extreme point. As shown in fig. 4, the obtained gaze sample trajectory is shown, where a black point A, B, C, D is an extreme point of a low-frequency trajectory of the gaze sample obtained after wavelet transform processing is performed on the gaze sample trajectory, and the gaze sample trajectory is finally divided into 5 initial trajectory segments.
In this optional embodiment, after all the gaze sample trajectories in the gaze positive sample trajectory set and the gaze negative sample trajectory set are segmented, all trajectory initial segments corresponding to the obtained gaze sample trajectories in the gaze positive sample trajectory set and the gaze negative sample trajectory set are used as a positive trajectory initial segment set and a negative trajectory initial segment set in the present solution, and the positive trajectory initial segment set and the negative trajectory initial segment set are used as trajectory initial segment sets.
In an alternative embodiment, the screening unit 112 is configured to screen the initial segment set of tracks to obtain a positive track valid segment set and a negative track valid segment set.
In an optional embodiment, the screening the initial track segment set to obtain a positive track valid segment set and a negative track valid segment set includes:
calculating the segment similarity of a target positive track initial segment and each positive track initial segment in the positive track initial segment set, wherein the target positive track initial segment is any one of the positive track initial segment set;
obtaining a positive track similar segment set corresponding to the target positive track initial segment based on a preset threshold and the segment similarity;
traversing each positive track initial segment in the positive track initial segment set to obtain a positive track similar segment set corresponding to each positive track initial segment;
calculating the segment similarity of a target negative track initial segment and each negative track initial segment in the negative track initial segment set, wherein the target negative track initial segment is any one of the negative track initial segment set;
obtaining a negative track similar segment set corresponding to the target negative track initial segment based on a preset threshold and the segment similarity;
traversing each negative track initial segment in the negative track initial segment set to obtain a negative track similar segment set corresponding to each negative track initial segment;
and respectively screening all the positive track similar fragment sets and all the negative track similar fragment sets to obtain a positive track effective fragment set and a negative track effective fragment set.
In this optional embodiment, the segment similarity between the target positive track initial segment and each positive track initial segment in the positive track initial segment set may be calculated according to a Dynamic Time Warping (DTW) algorithm, where the target positive track initial segment is any one of the positive track initial segment sets.
In this alternative embodiment, the DTW algorithm is a method for measuring similarity between two time sequences, and since different time sequences may only have a displacement in the time axis, the DTW algorithm may calculate the similarity between the two time sequences by extending and shortening the time sequences, compared to the conventional method for calculating the euclidean distance between the two time sequences. As shown in fig. 5, the upper and lower solid lines represent a target positive track initial segment and any one of the positive track initial segments in the positive track initial segment set except the target positive track initial segment, respectively, the dotted line between the two positive track initial segments represents the similar points between the two positive track initial segments, and the DTW algorithm measures the similarity between the two positive track initial segments as the segment similarity by calculating the sum of the distances between all the similar points. Because the DTW algorithm has various variants, the specific calculation process in the scheme can be consistent with the calculation process of the conventional DTW algorithm.
In this optional embodiment, after obtaining the segment similarity between the target positive track initial segment and each of the other positive track initial segments, filtering all the segment similarities corresponding to the target positive track initial segment based on a preset threshold, and taking the positive track initial segment corresponding to the segment similarity greater than the preset threshold as the positive track similar segment corresponding to the target positive track initial segment; and filtering the positive track initial segment corresponding to the segment similarity not greater than the preset threshold, and finally taking all positive track similar segments corresponding to the target positive track initial segment as a positive track similar segment set. The preset threshold of the segment similarity may be 0.6, that is, the positive track initial segment corresponding to the segment similarity greater than 0.6 is used as the positive track similar segment corresponding to the target positive track initial segment; and filtering out the positive track initial segment corresponding to the segment similarity not greater than 0.6.
In this alternative embodiment, a positive trace similar segment set corresponding to each positive trace initial segment may be obtained by traversing each positive trace initial segment in the positive trace initial segment set. And calculating to obtain a negative track similar segment set corresponding to each negative track initial segment, wherein the process is consistent with the acquisition process of the positive track similar segment set.
In this alternative embodiment, all positive trajectory similar segment sets and all negative trajectory similar segment sets may be sorted according to the sequence of the number of segments from large to small, so as to obtain a positive trajectory sorting result and a negative trajectory sorting result. Then, a positive track similar segment set and a negative track similar segment set ranked in a previous preset order are respectively selected from the positive track sorting result and the negative track sorting result, for example, the previous preset order may be 50, that is, 50 positive track similar segment sets and 50 negative track similar segment sets are finally obtained according to the number of the segments. In the scheme, each positive track similar segment set obtained through screening is used as a positive track effective segment set, and each negative track similar segment set is used as a negative track effective segment set.
In an alternative embodiment the construction unit 113 is adapted to construct a gaze track fraud identifier based on said set of positive track valid segments and said set of negative track valid segments.
In an optional embodiment, said constructing a gaze track fraud identifier based on said set of positive track valid segments and said set of negative track valid segments comprises:
extracting the longest common sequence of all the segments in each positive track effective segment set;
weighting and summing the longest public sequence of all positive track effective fragment sets to obtain a positive track reference template;
extracting the longest common sequence of all the segments in each negative track effective segment set;
carrying out weighted summation on the longest common sequence of all negative track effective fragment sets to obtain a negative track reference template;
and taking the positive track reference template and the negative track reference template as gaze track fraud identifiers.
In this optional embodiment, since a plurality of track valid segments similar to each other exist in each positive track valid segment set and negative track valid segment set, and the track valid segments are not completely the same, in order to obtain the common characteristics of all track valid segments in each positive track valid segment set and negative track valid segment set, the longest common sequence of all track valid segments in each positive track valid segment set and negative track valid segment set is extracted by a Longest Common Subsequence (LCS) algorithm in the present solution.
In this alternative embodiment, the lcs algorithm is used to find the longest subsequence of all sequences in a sequence set, such as str1= "ABCBDCA", str2= "DABDA", one common subsequence of both being AA and the longest common subsequence being ABDA. Wherein the characters in the common subsequence need not be contiguous in the original sequence.
In this optional embodiment, after the longest common subsequences of all segments in each positive track effective segment set and each negative track effective segment set are extracted, weights may be given to the corresponding longest common subsequences according to the number of track effective segments included in the positive track effective segment set or the negative track effective segment set where each longest common subsequence is located, and weighted summation is performed on the longest common subsequences of all positive track effective segment sets and all negative track effective segment sets where the weights are obtained, so as to obtain a positive track reference template and a negative track reference template respectively.
In this optional embodiment, taking the acquisition of the positive track reference templates as an example, a specific process of giving a weight to the longest common subsequence corresponding to each positive track reference template is as follows: the total number of positive track effective fragments included in a positive track effective fragment set in which each longest public subsequence is located is counted, all the total numbers are summed to obtain the total number of fragments, and the ratio of the total number of positive track effective fragments corresponding to each longest public subsequence to the total number of fragments is used as the weight of the corresponding longest public subsequence.
In this optional embodiment, the process of obtaining the negative track reference template in the present embodiment is the same as the process of obtaining the positive track reference template, and the finally obtained positive track reference template and the negative track reference template are used together as a gaze track fraud identifier for performing anti-fraud detection on a gaze track to be detected.
In an optional embodiment, the detection unit 114 is configured to perform anti-fraud detection on the target gaze track to be detected based on the gaze track fraud identifier to obtain a detection result.
In an optional embodiment, the performing, based on the gaze track fraud identifier, fraud detection on the target gaze track to be detected to obtain a detection result includes:
respectively calculating the similarity between a target gaze track to be detected and a positive track reference template and a negative track reference template in the gaze track fraud recognizer to obtain positive track similarity and negative track similarity;
if the positive track similarity is greater than the negative track similarity, the target gaze track to be detected has a fraud risk, and the detection result is a fraud high risk;
and if the positive track similarity is not greater than the negative track similarity, the target gaze track to be detected does not have fraud risk, and the detection result is fraud low risk.
In this optional embodiment, the obtained positive track similarity and the negative track similarity are compared, and if the positive track similarity is greater than the negative track similarity, it indicates that the target gaze track to be detected is closer to the positive track reference template in the gaze track fraud identifier, and has a fraud risk, and the detection result is a fraud high risk; and if the positive track similarity is not greater than the negative track similarity, the target gaze track to be detected is closer to the negative track reference template in the gaze track fraud identifier, and the target gaze track to be detected does not have fraud risk, and the detection result is low risk of fraud.
In an alternative embodiment, the gaze track fraud identifier may be updated based on the detection results to improve the detection accuracy of the gaze track fraud identifier.
In this optional embodiment, the gaze track fraud identifier may be updated based on the detection result, and if the detection result is a high risk of fraud, the target gaze track corresponding to the high risk of fraud is marked as a gaze positive sample and added to the gaze positive sample track set, so that a new positive track reference template is generated through a series of processing procedures shown in this scheme to update the gaze track fraud identifier; and if the detection result is fraud low risk, marking the target gaze track corresponding to the fraud high risk as a gaze negative sample, adding the gaze negative sample track set into the gaze negative sample track set, and finally generating a new negative track reference template to update the gaze track fraud identifier.
According to the technical scheme, the positive and negative eye sight sample tracks are extracted from the obtained eye sight video data and are combined after the positive and negative eye sight sample tracks are segmented, so that the eye sight track fraud recognizer capable of detecting whether the eye sight tracks contain fraud information is obtained, and the accuracy of recognizing the fraud risk of the time sequence eye sight tracks is effectively improved.
Please refer to fig. 3, which is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure. The electronic device 1 comprises a memory 12 and a processor 13. The memory 12 is used for storing computer readable instructions, and the processor 13 is used for executing the computer readable instructions stored in the memory to implement the artificial intelligence based gaze track anti-fraud method according to any of the above embodiments.
In an alternative embodiment, the electronic device 1 further comprises a bus, a computer program stored in said memory 12 and executable on said processor 13, such as an artificial intelligence based gaze track anti-fraud program.
Fig. 3 shows only the electronic device 1 with the memory 12 and the processor 13, and it will be understood by those skilled in the art that the structure shown in fig. 3 does not constitute a limitation of the electronic device 1, and may comprise fewer or more components than shown, or a combination of certain components, or a different arrangement of components.
In conjunction with fig. 1, the memory 12 in the electronic device 1 stores a plurality of computer-readable instructions to implement an artificial intelligence based gaze track anti-fraud method, and the processor 13 is capable of executing the plurality of instructions to implement:
acquiring eye video data to generate an eye sample track set, wherein the eye sample track set comprises an eye positive sample track set and an eye negative sample track set;
segmenting all gaze sample tracks in the gaze sample track set to obtain a track initial segment set, wherein the track initial segment set comprises a positive track initial segment set and a negative track initial segment set;
screening the initial track fragment set to obtain a positive track effective fragment set and a negative track effective fragment set;
constructing a gaze track fraud recognizer based on the positive track valid segment set and the negative track valid segment set;
and carrying out anti-fraud detection on the target gaze track to be detected based on the gaze track fraud identifier to obtain a detection result.
Specifically, the processor 13 may refer to the description of the relevant steps in the embodiment corresponding to fig. 1 for a specific implementation method of the instruction, which is not described herein again.
It will be appreciated by those skilled in the art that the schematic diagram is merely an example of the electronic device 1, and does not constitute a limitation to the electronic device 1, the electronic device 1 may have a bus-type structure or a star-shaped structure, and the electronic device 1 may further include more or less hardware or software than that shown in the figure, or different component arrangements, for example, the electronic device 1 may further include an input and output device, a network access device, and the like.
It should be noted that the electronic device 1 is only an example, and other existing or future electronic products, such as those that may be adapted to the present application, should also be included in the scope of protection of the present application, and are included by reference.
The processor 13 may be composed of an integrated circuit in some embodiments, for example, a single packaged integrated circuit, or may be composed of a plurality of integrated circuits packaged with the same or different functions, including one or more Central Processing Units (CPUs), microprocessors, digital Processing chips, graphics processors, and combinations of various control chips. The processor 13 is a Control Unit (Control Unit) of the electronic device 1, connects various components of the electronic device 1 by using various interfaces and lines, and executes various functions and processes data of the electronic device 1 by running or executing programs or modules stored in the memory 12 (for example, executing an artificial intelligence-based eye tracking anti-cheating program, etc.), and calling data stored in the memory 12.
The processor 13 executes an operating system of the electronic device 1 and various installed application programs. The processor 13 executes the application program to implement the steps in each artificial intelligence based gaze track anti-fraud method embodiment described above, such as the steps shown in fig. 1.
Illustratively, the computer program may be partitioned into one or more modules/units, which are stored in the memory 12 and executed by the processor 13 to accomplish the present application. The one or more modules/units may be a series of computer-readable instruction segments capable of performing certain functions, which are used to describe the execution of the computer program in the electronic device 1. For example, the computer program may be divided into an acquisition unit 110, a segmentation unit 111, a screening unit 112, a construction unit 113, a detection unit 114.
The integrated unit implemented in the form of a software functional module may be stored in a computer-readable storage medium. The software functional module is stored in a storage medium and includes several instructions to enable a computer device (which may be a personal computer, a computer device, or a network device) or a processor (processor) to execute the artificial intelligence based gaze track anti-fraud method according to the embodiments of the present application.
The integrated modules/units of the electronic device 1 may be stored in a computer-readable storage medium if they are implemented in the form of software functional units and sold or used as separate products. Based on such understanding, all or part of the flow in the method of the embodiments described above may be implemented by a computer program, which may be stored in a computer readable storage medium, and when the computer program is executed by a processor, the steps of the embodiments of the methods described above may be implemented.
Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U.S. disk, removable hard disk, magnetic diskette, optical disk, computer Memory, read-Only Memory (ROM), random access Memory and other Memory, etc.
Further, the computer-readable storage medium may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the blockchain node, and the like.
The block chain referred by the application is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, a consensus mechanism, an encryption algorithm and the like. A block chain (Blockchain), which is essentially a decentralized database, is a series of data blocks associated by using a cryptographic method, and each data block contains information of a batch of network transactions, so as to verify the validity (anti-counterfeiting) of the information and generate a next block. The blockchain may include a blockchain underlying platform, a platform product service layer, an application service layer, and the like.
The bus may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one arrow is shown in FIG. 3, but this does not indicate only one bus or one type of bus. The bus is arranged to enable connection communication between the memory 12 and at least one processor 13 or the like.
An embodiment of the present application further provides a computer-readable storage medium (not shown), where the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions are executed by a processor in an electronic device to implement the artificial intelligence based gaze track anti-fraud method according to any of the above embodiments.
In the several embodiments provided in the present application, it should be understood that the disclosed system, apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is only one logical functional division, and other divisions may be realized in practice.
The modules described as separate parts may or may not be physically separate, and parts displayed as modules may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, functional modules in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional module.
Furthermore, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. A plurality of units or means recited in the specification may also be implemented by one unit or means through software or hardware. The terms first, second, etc. are used to denote names, but not any particular order.
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present application and not for limiting, and although the present application is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made on the technical solutions of the present application without departing from the spirit and scope of the technical solutions of the present application.
Claims (10)
1. An artificial intelligence-based gaze track anti-fraud method is characterized by comprising the following steps:
acquiring eye video data to generate an eye sample track set, wherein the eye sample track set comprises an eye positive sample track set and an eye negative sample track set;
segmenting all gaze sample tracks in the gaze sample track set to obtain a track initial segment set, wherein the track initial segment set comprises a positive track initial segment set and a negative track initial segment set;
screening the initial track fragment set to obtain a positive track effective fragment set and a negative track effective fragment set;
constructing a gaze track fraud recognizer based on the positive track valid segment set and the negative track valid segment set;
and carrying out anti-fraud detection on the target gaze track to be detected based on the gaze track fraud identifier to obtain a detection result.
2. The artificial intelligence based gaze track antifraud method of claim 1, wherein the obtaining video data to generate a gaze sample track set comprises:
collecting eye sight video data;
carrying out positive and negative sample marking on all gaze video data according to a preset mode, taking all gaze video data with positive sample marks as a gaze positive sample data set, and taking all gaze video data with negative sample marks as a gaze negative sample data set;
generating eye gaze positive sample tracks based on the eye gaze video data in the eye gaze positive sample data set, and taking all eye gaze positive sample tracks as an eye gaze positive sample track set;
generating gaze negative sample tracks based on gaze video data in the gaze negative sample data set, and taking all gaze negative sample tracks as a gaze negative sample track set;
and taking the gaze positive sample trajectory set and the gaze negative sample trajectory set as gaze sample trajectory sets.
3. The artificial intelligence based gaze track anti-fraud method of claim 1, wherein the slicing all gaze sample tracks in the gaze sample track set to obtain an initial segment set of tracks comprises:
extracting low-frequency information of each gaze sample track in the gaze sample track set to obtain a gaze sample low-frequency track;
calculating extreme points of the low-frequency track of the gaze sample, and segmenting the low-frequency track of the gaze sample according to all the extreme points to obtain a plurality of track initial segments of the low-frequency track of the gaze sample;
respectively taking all track initial segments corresponding to the gaze sample tracks in the gaze positive sample track set and the gaze negative sample track set as a positive track initial segment set and a negative track initial segment set;
and taking the positive track initial segment set and the negative track initial segment set as track initial segment sets.
4. The artificial intelligence based gaze track anti-fraud method of claim 1, wherein the screening the initial set of track segments to obtain a positive set of track valid segments and a negative set of track valid segments comprises:
calculating the segment similarity of a target positive track initial segment and each positive track initial segment in the positive track initial segment set, wherein the target positive track initial segment is any one of the positive track initial segment set;
obtaining a positive track similar segment set corresponding to the target positive track initial segment based on a preset threshold and the segment similarity;
traversing each positive track initial segment in the positive track initial segment set to obtain a positive track similar segment set corresponding to each positive track initial segment;
calculating the segment similarity of a target negative track initial segment and each negative track initial segment in the negative track initial segment set, wherein the target negative track initial segment is any one of the negative track initial segment set;
obtaining a negative track similar segment set corresponding to the target negative track initial segment based on a preset threshold and the segment similarity;
traversing each negative track initial segment in the negative track initial segment set to obtain a negative track similar segment set corresponding to each negative track initial segment;
and respectively screening all the positive track similar fragment sets and all the negative track similar fragment sets to obtain a positive track effective fragment set and a negative track effective fragment set.
5. The artificial intelligence based gaze track anti-fraud method of claim 4, wherein the screening all positive track similar segment sets and all negative track similar segment sets to obtain positive track valid segment sets and negative track valid segment sets respectively comprises:
respectively counting the number of the segments included in each positive track similar segment set and each negative track similar segment set;
sequencing all positive track similar segment sets and all negative track similar segment sets respectively according to the sequence of the segment numbers from large to small to obtain positive track sequencing results and negative track sequencing results;
and respectively screening all positive track similar segment sets and all negative track similar segment sets based on the positive track sorting result and the negative track sorting result to obtain a plurality of positive track effective segment sets and negative track effective segment sets with the same quantity.
6. The artificial intelligence based gaze track fraud prevention method of claim 1, wherein said constructing a gaze track fraud identifier based on the positive track valid segment set and the negative track valid segment set comprises:
extracting the longest public sequence of all the segments in each positive track effective segment set;
weighting and summing the longest common sequence of all positive track effective fragment sets to obtain a positive track reference template;
extracting the longest common sequence of all the segments in each negative track effective segment set;
carrying out weighted summation on the longest common sequence of all negative track effective fragment sets to obtain a negative track reference template;
and taking the positive track reference template and the negative track reference template as eye track fraud identifiers.
7. The artificial intelligence based gaze track anti-fraud method according to claim 1, wherein the anti-fraud detection of the target gaze track to be detected based on the gaze track fraud identifier to obtain the detection result comprises:
respectively calculating the similarity between a target gaze track to be detected and a positive track reference template and a negative track reference template in the gaze track fraud identifier to obtain a positive track similarity and a negative track similarity;
if the positive track similarity is greater than the negative track similarity, the target gaze track to be detected has a fraud risk, and the detection result is a fraud high risk;
and if the positive track similarity is not greater than the negative track similarity, the target gaze track to be detected does not have fraud risk, and the detection result is fraud low risk.
8. An artificial intelligence based gaze track anti-fraud device, the device comprising:
an obtaining unit, configured to obtain eye video data to generate a eye sample trajectory set, where the eye sample trajectory set includes an eye positive sample trajectory set and an eye negative sample trajectory set;
the segmentation unit is used for segmenting all gaze sample tracks in the gaze sample track set to obtain a track initial segment set, wherein the track initial segment set comprises a positive track initial segment set and a negative track initial segment set;
the screening unit is used for screening the track initial segment set to obtain a positive track effective segment set and a negative track effective segment set;
the construction unit is used for constructing a gaze track fraud recognizer based on the positive track effective segment set and the negative track effective segment set;
and the detection unit is used for carrying out anti-fraud detection on the target gaze track to be detected based on the gaze track fraud identifier to obtain a detection result.
9. An electronic device, characterized in that the electronic device comprises:
a memory storing computer readable instructions; and
a processor executing computer readable instructions stored in the memory to implement the artificial intelligence based gaze track anti-fraud method of any of claims 1 to 7.
10. A computer-readable storage medium having computer-readable instructions stored thereon which, when executed by a processor, implement the artificial intelligence based gaze track antifraud method of any of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210987439.2A CN115346263A (en) | 2022-08-17 | 2022-08-17 | Eye track anti-fraud method, device, equipment and medium based on artificial intelligence |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210987439.2A CN115346263A (en) | 2022-08-17 | 2022-08-17 | Eye track anti-fraud method, device, equipment and medium based on artificial intelligence |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115346263A true CN115346263A (en) | 2022-11-15 |
Family
ID=83951748
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210987439.2A Pending CN115346263A (en) | 2022-08-17 | 2022-08-17 | Eye track anti-fraud method, device, equipment and medium based on artificial intelligence |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115346263A (en) |
-
2022
- 2022-08-17 CN CN202210987439.2A patent/CN115346263A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CA2791597C (en) | Biometric training and matching engine | |
CN112669138B (en) | Data processing method and related equipment | |
CN111860377A (en) | Live broadcast method and device based on artificial intelligence, electronic equipment and storage medium | |
CN112232203B (en) | Pedestrian recognition method and device, electronic equipment and storage medium | |
CN111369801B (en) | Vehicle identification method, device, equipment and storage medium | |
CN111798047A (en) | Wind control prediction method and device, electronic equipment and storage medium | |
CN111709382A (en) | Human body trajectory processing method and device, computer storage medium and electronic equipment | |
CN110827159B (en) | Financial medical insurance fraud early warning method, device and terminal based on relation diagram | |
CN113470775B (en) | Information acquisition method, device, equipment and storage medium | |
Wiesner et al. | Dataset of digitized RACs and their rarity score analysis for strengthening shoeprint evidence | |
CN117611569A (en) | Vehicle fascia detection method, device, equipment and medium based on artificial intelligence | |
CN115795466B (en) | Malicious software organization identification method and device | |
Wang et al. | Marathon athletes number recognition model with compound deep neural network | |
CN115346263A (en) | Eye track anti-fraud method, device, equipment and medium based on artificial intelligence | |
CN115131564A (en) | Vehicle component damage detection method based on artificial intelligence and related equipment | |
CN113349792B (en) | Method, apparatus, device and medium for classifying multi-lead electrocardiosignal | |
CN115293809A (en) | Typhoon and rainstorm risk rating method based on artificial intelligence and related equipment | |
CN112200472B (en) | Method and device for detecting accompanying vehicle, electronic equipment and storage medium | |
CN115022014B (en) | Login risk identification method, device, equipment and storage medium | |
CN115001802B (en) | Account abnormal login detection method based on shared screen and related equipment | |
CN114972883B (en) | Target detection sample generation method based on artificial intelligence and related equipment | |
Kim et al. | Beyond visual inspection: capturing neighborhood dynamics with historical Google Street View and deep learning-based semantic segmentation | |
CN110020663B (en) | Risk label processing method and device and electronic equipment | |
CN115130610A (en) | Classification model construction method based on artificial intelligence and related equipment | |
CN114625442A (en) | Cold start recommendation method and device, electronic equipment and readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination |