CN110598042A - Incremental update-based video structured real-time updating method and system - Google Patents
Incremental update-based video structured real-time updating method and system Download PDFInfo
- Publication number
- CN110598042A CN110598042A CN201910806616.0A CN201910806616A CN110598042A CN 110598042 A CN110598042 A CN 110598042A CN 201910806616 A CN201910806616 A CN 201910806616A CN 110598042 A CN110598042 A CN 110598042A
- Authority
- CN
- China
- Prior art keywords
- real
- time
- vectors
- clustering
- updating
- 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
- 238000000034 method Methods 0.000 title claims abstract description 31
- 239000013598 vector Substances 0.000 claims abstract description 72
- 238000000605 extraction Methods 0.000 claims abstract description 26
- 238000001514 detection method Methods 0.000 claims abstract description 17
- 238000012216 screening Methods 0.000 claims abstract description 12
- 238000001914 filtration Methods 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 abstract description 6
- 238000005516 engineering process Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 5
- 238000013473 artificial intelligence Methods 0.000 description 3
- 230000002354 daily effect Effects 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009960 carding Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/71—Indexing; Data structures therefor; Storage structures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/75—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7837—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using objects detected or recognised in the video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7837—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using objects detected or recognised in the video content
- G06F16/784—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using objects detected or recognised in the video content the detected or recognised objects being people
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Library & Information Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a video structured real-time updating method and a system based on incremental updating, wherein the method comprises the following steps of: extracting information through key frame screening, target detection, label identification and feature extraction, abstracting a picture into high-dimensional vectors and storing the high-dimensional vectors; classifying the multi-level cluster indexes: clustering all stored vectors according to classification, establishing indexes, and clustering the results again; pre-establishing a real-time index: according to indexes established by classified multi-level clustering, the indexes established by the previous data of each camera are used as reference classes for real-time classification of each camera; vector real-time updating: judging and classifying the vectors generated in real time, and writing the vectors into the most similar directory; and (4) incremental updating: and re-clustering and sorting the added real-time vectors at regular time. The invention can greatly improve the use efficiency of the monitoring system and the daily work efficiency of the user.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a video structured real-time updating method and system based on incremental updating.
Background
Some users often need to check video monitoring contents in daily work, but the massive cameras bring great challenges to the work of the users, and the users often need to watch and screen video data of a large number of cameras for a long time and in high strength during work, so that the time and labor are consumed, and clues can be missed. Under the background, a video structured real-time updating system integrating artificial intelligence technologies such as target detection, feature extraction and label extraction and a data increment updating algorithm is required to be carried, monitored video contents are extracted and combed for a user, information retrieval can be carried out through an external searching interface, and therefore the use efficiency of the monitoring system and the daily work efficiency of the user are greatly improved.
An effective solution to the problems in the related art has not been proposed yet.
Disclosure of Invention
Aiming at the technical problems in the related art, the invention provides a video structured real-time updating method and system based on incremental updating, which can be used for summarizing and extracting real-time information by using various artificial intelligence algorithms and clustering the extracted data information in real time to store the data information in an indexed structure, so that the data can be searched in real time, and the content understanding, the information extraction, the structured storage and the real-time updating of the monitoring camera video can be realized.
In order to achieve the technical purpose, the technical scheme of the invention is realized as follows:
a video structured real-time updating method based on incremental updating comprises the following steps:
information extraction: extracting information through key frame screening, target detection, label identification and feature extraction, abstracting a picture into high-dimensional vectors and storing the high-dimensional vectors;
classifying the multi-level cluster indexes: clustering all stored vectors according to classification, establishing indexes, and clustering the results again;
pre-establishing a real-time index: according to indexes established by classified multi-level clustering, the indexes established by the previous data of each camera are used as reference classes for real-time classification of each camera;
vector real-time updating: judging and classifying the vectors generated in real time, and writing the vectors into the most similar directory;
and (4) incremental updating: and re-clustering and sorting the added real-time vectors at regular time.
Further, the key frame screening is to screen a picture obtained from an uploaded video or a camera frame cut, remove repeated frames in adjacent frames, and only retain key frames with more information.
Further, the target detection is to detect and segment a large image obtained by a cut frame to obtain a person/vehicle target and related information in the image.
Further, the tag identification is to perform tag identification on the detected person/vehicle target to obtain a person/vehicle attribute tag, write the person/vehicle attribute tag into a database, and perform structured storage.
Further, the feature extraction is to perform feature extraction on the pictures to be searched, and abstract each picture into a high-dimensional vector.
A video structured real-time updating system based on incremental updating comprises the following modules:
the information extraction module: extracting information through a key frame screening module, a target detection module, a tag identification module and a feature extraction module, abstracting a picture into high-dimensional vectors and storing the high-dimensional vectors;
the classification multistage clustering index module: clustering all stored vectors according to classification, establishing indexes, and clustering the results again;
a real-time index pre-establishing module: according to indexes established by classified multi-level clustering, the indexes established by the previous data of each camera are used as reference classes for real-time classification of each camera;
the vector real-time updating module: judging and classifying the vectors generated in real time, and writing the vectors into the most similar directory;
an incremental update module: and re-clustering and sorting the added real-time vectors at regular time.
Further, the key frame screening module screens pictures obtained from uploaded videos or camera frame cutting, removes repeated frames in adjacent frames, and only retains key frames with more information.
Further, the target detection module detects and segments a large image obtained by a frame cutting, and obtains a person/vehicle target and related information in the image.
Further, the tag identification module is used for carrying out tag identification on the detected person/vehicle target to obtain a person/vehicle attribute tag, writing the person/vehicle attribute tag into a database and carrying out structured storage.
Further, the feature extraction module is used for extracting features of pictures to be searched, and abstracting each picture into a high-dimensional vector.
The invention has the beneficial effects that:
1. the utilization rate of the mass monitoring cameras is improved, and key information in the cameras is intelligently acquired and combed.
2. The efficiency of relevant work is promoted, the time and the energy that the user need expend when watching the surveillance video and looking for information are greatly reduced, and the manpower is replaced with the mechanical computing power, so that the people can make decisions.
3. The intelligent degree is high, various information which may be needed by a user can be extracted from the video, and potential and actual requirements are met.
4. The system is additionally provided with a video analysis function, the process of extracting key information from videos watched by manpower can be compressed by tens of times for videos with any standard format, the information is efficiently and accurately combed, and the videos can be searched through organized storage.
5. The real-time updating of the camera information to the library is realized, the real-time searching can be realized, the timeliness of the information is well utilized and protected, and the working efficiency can be greatly improved.
6. An increment updating technology is designed, data sorting time is greatly shortened, data can keep a good storage and index structure at any time, and the accuracy and efficiency of searching are improved.
7. The method constructs multi-level indexes for the mass data, optimizes a data storage structure, fully exerts the advantages of a parallel retrieval algorithm and realizes efficient data sorting and searching.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a basic flowchart of a video structured real-time update method based on incremental update according to an embodiment of the present invention.
Fig. 2 is a detailed data processing flow diagram of a video structured real-time update system based on incremental update according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments that can be derived by one of ordinary skill in the art from the embodiments given herein are intended to be within the scope of the present invention.
As shown in fig. 1, a method for updating a video structure based on incremental update in real time according to an embodiment of the present invention includes the following steps:
information extraction: extracting information through key frame screening, target detection, label identification and feature extraction, abstracting a picture into high-dimensional vectors and storing the high-dimensional vectors;
classifying the multi-level cluster indexes: clustering all stored vectors according to classification, establishing indexes, and clustering the results again;
pre-establishing a real-time index: according to indexes established by classified multi-level clustering, the indexes established by the previous data of each camera are used as reference classes for real-time classification of each camera;
vector real-time updating: judging and classifying the vectors generated in real time, and writing the vectors into the most similar directory;
and (4) incremental updating: and re-clustering and sorting the added real-time vectors at regular time.
In the preferred embodiment, the key frame screening is to screen a picture obtained from an uploaded video or a camera frame cut, remove repeated frames in adjacent frames, and only retain key frames with more information.
In the preferred embodiment, the target detection is to detect and segment a large image obtained by a frame to obtain a human/vehicle target and related information in the image.
In the preferred embodiment, the tag identification is to perform tag identification on the detected person/vehicle object, obtain a person/vehicle attribute tag, write the person/vehicle attribute tag into a database, and perform structured storage.
In the preferred embodiment, the feature extraction is to extract features of the pictures to be searched, and each picture is abstracted into a high-dimensional vector.
As shown in fig. 2, according to the video structured real-time update system based on incremental update according to the embodiment of the present invention, the working principle and the flow of each module are as follows:
information extraction part
Video crawler/camera data interfacing
First is the source of the data: the camera meets the national standard GB 28181. The system is connected with a camera meeting the standard to obtain video stream data, and operations such as frame cutting to obtain screenshots, timed storage to obtain videos and the like can be executed from the video stream. The pictures and videos are stored on the data hard disk, information metadata related to the videos are written into a table of a database, and the metadata related to the pictures are written into a message queue for consumption and use of a following module.
Video upload
The user can upload videos in standard formats as input to the system, and the system can analyze and structure the video contents at high speed and can search the video contents. Compared with a manual video watching process, the process is improved by several times or even tens of times, and the offline video can be analyzed and data can be extracted, so that the data can be updated to the library in real time.
Key frame screening
The method mainly screens pictures obtained from uploaded videos or camera frame cutting, removes repeated frames in adjacent frames, and only reserves key frames with more information. Before pictures are obtained by frame cutting of a camera and relevant information of the pictures is stored and transmitted, the pictures need to be screened, and one or more frames are reserved in continuous multi-frame repeated or similar pictures, so that a large amount of repeated and redundant data is removed. When the key frame screening is carried out on the picture, the previous picture and the next picture of the picture are taken as references, a difference value between two frames is calculated by using a three-frame difference method, a standard difference of the difference value on the whole picture is calculated, and when the standard difference reaches a preset threshold value, the picture and the previous and next continuous frames are considered to have larger difference and belong to the key frame, so that the difference is reserved.
Target detection
The method mainly comprises the steps of detecting and segmenting a large image obtained by a cut frame, and obtaining a person/vehicle target and related information in the image. The screened pictures are stored in a hard disk and relevant information (path, time and the like) is written into a message queue, a target detection algorithm module is used as a consumer to consume messages in the queue in real time, corresponding pictures are read from the hard disk according to the path and sent into a pre-trained target detection model to obtain a person/vehicle target list in each picture and a corresponding position in the picture, then original pictures are cut, the cut person/vehicle small pictures are stored on a data hard disk, and corresponding metadata are sent to a tag identification algorithm module to carry out tag identification.
Label identification
The method mainly comprises the steps of carrying out label identification on a detected person/vehicle target, obtaining a person/vehicle attribute label (including target attribute, picture time, place and the like), writing the person/vehicle attribute label into a database, and carrying out structured storage.
1. And (3) taking the human/vehicle target output by the target detection algorithm model as the input of the label identification model, and identifying the attribute corresponding to the human/vehicle target by using the pre-trained model, such as: (car) color, (car) primary and secondary driving safety belts, (car) primary and secondary driving sun shield plates, (car) whether license plate is shielded, (man) satchel, (man) backpack, (man) carry-on object, (man) jacket style color, (man) trousers style color, (man) gender, (man) glasses, (man) mask, and the like.
2. After the algorithm detects the people/vehicles and the corresponding attributes, the information of each target, the target labels, the corresponding big pictures and the like is written into a corresponding table of the ES database and can be searched and inquired.
Feature extraction/picture vectorization
The method mainly comprises the steps of extracting features of pictures to be searched, abstracting each picture into a high-dimensional vector (corresponding dimensions of different types of pictures are different), and using the high-dimensional vector for establishing indexes and searching comparison operation. And (3) using a pre-trained vectorization model to extract the features of all the pictures such as a cut-frame big picture/a human small picture/a car small picture/a big picture segmentation picture and the like, converting the pictures into high-dimensional vectors in a one-to-one correspondence manner, writing the vectors and attribute information into a message queue, and transmitting the vectors and the attribute information to a real-time clustering module for use.
Two, picture vector clustering storage part
Categorizing the multilevel clustering index (! important)
The method mainly comprises the steps of clustering all stored vectors according to the classification of time and camera places, establishing indexes, clustering results again, and reducing the data volume during searching so as to realize efficient and accurate searching. Vectors of different classes (big picture, car, person, segmentation picture, etc.) are clustered individually.
1. The data are divided according to the day and the camera and are clustered by using algorithms respectively, so that the data with different conditions are ensured to have independence respectively, a user can add conditions such as time, place and the like during searching, and higher efficiency and accuracy can be realized for clue searching in a limited range.
2. For the data volume of billions or even billions supported by the system, the number of indexes obtained after the initial clustering is still in the level of ten million. On the basis, clustering is carried out on the ten-million indexes again to obtain second-level indexes, and the number of the second-level indexes is compressed to one hundred thousand levels. And for millions and below of data volumes, the search operation time is millisecond.
3. For the multi-level index obtained by the method, the efficiency of searching by using an external search interface is improved by times.
Real-time index pre-building
The method mainly comprises the steps of establishing indexes according to classified multistage clustering, and taking the indexes established by previous data of each camera as a reference class for real-time classification of each camera. Because the system needs to process the vectors in real time and classify the vectors to update in real time, an index directory needs to be pre-established to process the real-time vectors, the most similar index is obtained by traversing the real-time index directory once, and the vectors are written into the corresponding position of the hard disk. The vectors thus processed are already stored in a position that can be searched in real time.
Vector real-time update
The method mainly comprises the steps of judging and classifying vectors generated in real time, writing the vectors into the most similar directory, enabling the vectors to be searched immediately, and achieving real-time updating of data. And loading the pre-established real-time index into a video memory in advance by using the pre-established real-time index, traversing the indexes for all vectors generated in real time to find the most similar index, and writing the most similar index into a corresponding hard disk directory according to the time of the vectors and the attribute of a camera. Since all vectors in the directory are read according to the directory during searching, writing to the hard disk means that the vectors can be searched, and real-time updating of the data of the camera is realized.
Incremental update (delta update)
The method mainly includes that increased real-time vectors are re-clustered and sorted at regular time to guarantee accuracy and efficiency of searching, and meanwhile, data sorting time is greatly shortened by controlling an updated data range.
1. Generally, incremental updating can be performed when the system load is minimum every morning, all real-time vectors written in the previous day are clustered again, and the index directory with relatively more uniform distribution and good structure is sorted.
2. After the increment updating every day is finished, the secondary index is correspondingly regenerated to cover the old secondary index, and after the updating is finished, the data storage with a good searching structure is obtained and can be directly loaded during searching.
3. Through incremental updating, the data and the index have good structures all the time, good timeliness of human-vehicle data is guaranteed, real-time updating and searching can be achieved, and meanwhile, the human-vehicle data is organized and accurate.
Data retrieval method
1. For the label information corresponding to the picture, a search of keywords and characters can be provided.
2. For searching of pictures themselves, picture vectorization based picture searching can be provided based on an index structure of picture vectors. 3. Or combining and intersecting the result of the image search and the result of the label search, and providing the result as a result of the multi-condition search to the user.
In conclusion, the invention screens and eliminates the same targets (such as the same pictures when waiting for red light) of adjacent segments in the video by the key frame screening technology, only retains meaningful key frames, compresses the storage space and the search library in multiples, and realizes data duplication removal and redundancy removal; through a target detection technology, extraction of people/vehicle information appearing in a monitoring video is achieved; through a tag identification technology, the extracted person/vehicle is subjected to attribute identification, and the result is stored and can be searched, and can be searched based on texts or tags; by means of a feature extraction technology, the character/vehicle images are in one-to-one correspondence to a high-dimensional index containing all features of the character/vehicle images, so that image vectorization is realized, vectors are regarded as points in a high-dimensional space, and carding and searching are facilitated; through a classified multistage clustering index algorithm, multistage index establishment is realized by performing multistage clustering, structured storage of video data is realized, and searching can be performed in an index mode; by the index pre-establishing technology and the real-time classification technology, people and vehicles data obtained by the camera in real time are classified in real time, so that the people and vehicles data are updated and searched in real time, the problem of poor real-time performance of the data is solved, and the case handling efficiency is greatly improved; by the incremental updating (delta updating) technology, the data arrangement time is greatly shortened, the newly added data can be regularly arranged every day, the orderliness and the good structure of the data are kept, and the accuracy rate of clues and the utilization rate of the data are improved.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. A video structured real-time updating method based on incremental updating is characterized by comprising the following steps:
information extraction: extracting information through key frame screening, target detection, label identification and feature extraction, abstracting a picture into high-dimensional vectors and storing the high-dimensional vectors;
classifying the multi-level cluster indexes: clustering all stored vectors according to classification, establishing indexes, and clustering the results again;
pre-establishing a real-time index: according to indexes established by classified multi-level clustering, the indexes established by the previous data of each camera are used as reference classes for real-time classification of each camera;
vector real-time updating: judging and classifying the vectors generated in real time, and writing the vectors into the most similar directory;
and (4) incremental updating: and re-clustering and sorting the added real-time vectors at regular time.
2. The method according to claim 1, wherein the keyframe selection is performed by selecting a picture obtained from an uploaded video or a camera frame cut, removing a repeat frame from an adjacent frame, and only retaining a keyframe with more information.
3. The incremental update based video structured real-time update method as claimed in claim 1, wherein the target detection is to detect and segment a large image obtained from a slice frame to obtain a human/vehicle target and related information in the image.
4. The incremental update-based video structured real-time update method according to claim 1, wherein the tag identification is to perform tag identification on the detected person/vehicle object, obtain a person/vehicle attribute tag, write the person/vehicle attribute tag into a database, and perform structured storage.
5. The incremental update-based video structured real-time update method according to claim 1, wherein the feature extraction is to perform feature extraction on the pictures to be searched, and abstract each picture into a high-dimensional vector.
6. A video structured real-time updating system based on incremental updating is characterized by comprising the following modules:
the information extraction module: extracting information through a key frame screening module, a target detection module, a tag identification module and a feature extraction module, abstracting a picture into high-dimensional vectors and storing the high-dimensional vectors;
the classification multistage clustering index module: clustering all stored vectors according to classification, establishing indexes, and clustering the results again;
a real-time index pre-establishing module: according to indexes established by classified multi-level clustering, the indexes established by the previous data of each camera are used as reference classes for real-time classification of each camera;
the vector real-time updating module: judging and classifying the vectors generated in real time, and writing the vectors into the most similar directory;
an incremental update module: and re-clustering and sorting the added real-time vectors at regular time.
7. The video structured real-time updating system based on incremental updating of claim 6, wherein the key frame filtering module filters a picture obtained from an uploaded video or a camera frame cut, removes repeated frames in adjacent frames, and only retains key frames with more information.
8. The incremental update based video structured real-time update system as claimed in claim 6, wherein the object detection module detects and segments a large image obtained from a slice frame to obtain a human/vehicle object and related information in the image.
9. The incremental update-based video structured real-time update system according to claim 6, wherein the tag identification module performs tag identification on the detected human/vehicle object to obtain a human/vehicle attribute tag, and writes the human/vehicle attribute tag into the database for structured storage.
10. The incremental update based video structured real-time update system according to claim 6, wherein the feature extraction module performs feature extraction on the pictures to be searched, and abstracts each picture into a high-dimensional vector.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910806616.0A CN110598042A (en) | 2019-08-29 | 2019-08-29 | Incremental update-based video structured real-time updating method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910806616.0A CN110598042A (en) | 2019-08-29 | 2019-08-29 | Incremental update-based video structured real-time updating method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110598042A true CN110598042A (en) | 2019-12-20 |
Family
ID=68856240
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910806616.0A Pending CN110598042A (en) | 2019-08-29 | 2019-08-29 | Incremental update-based video structured real-time updating method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110598042A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111461044A (en) * | 2020-04-07 | 2020-07-28 | 北京波谱华光科技有限公司 | Video synthesis method, system and computer storage medium |
CN111586354A (en) * | 2020-04-28 | 2020-08-25 | 上海市保安服务(集团)有限公司 | Investigation system |
CN116320535A (en) * | 2023-04-14 | 2023-06-23 | 北京百度网讯科技有限公司 | Method, device, electronic equipment and storage medium for generating video |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101211355A (en) * | 2006-12-30 | 2008-07-02 | 中国科学院计算技术研究所 | Image inquiry method based on clustering |
CN102693299A (en) * | 2012-05-17 | 2012-09-26 | 西安交通大学 | System and method for parallel video copy detection |
US20130198783A1 (en) * | 2012-02-01 | 2013-08-01 | Futurewei Technologies, Inc. | System and Method for Organizing Multimedia Content |
CN106127114A (en) * | 2016-06-16 | 2016-11-16 | 北京数智源科技股份有限公司 | Intelligent video analysis method |
CN109743547A (en) * | 2018-12-27 | 2019-05-10 | 裴广华 | A kind of artificial intelligence security monitoring management system |
-
2019
- 2019-08-29 CN CN201910806616.0A patent/CN110598042A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101211355A (en) * | 2006-12-30 | 2008-07-02 | 中国科学院计算技术研究所 | Image inquiry method based on clustering |
US20130198783A1 (en) * | 2012-02-01 | 2013-08-01 | Futurewei Technologies, Inc. | System and Method for Organizing Multimedia Content |
CN102693299A (en) * | 2012-05-17 | 2012-09-26 | 西安交通大学 | System and method for parallel video copy detection |
CN106127114A (en) * | 2016-06-16 | 2016-11-16 | 北京数智源科技股份有限公司 | Intelligent video analysis method |
CN109743547A (en) * | 2018-12-27 | 2019-05-10 | 裴广华 | A kind of artificial intelligence security monitoring management system |
Non-Patent Citations (3)
Title |
---|
杨志荣 等: "用SOM聚类实现多级高维点数据索引", 《计算机研究与发展》 * |
袁文铎: "基于改进k-means的案例检索优化算法", 《计算机测量与控制》 * |
谢琳 等: "自然保护区海量视频数据快速分析与检索", 《计算机***应用》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111461044A (en) * | 2020-04-07 | 2020-07-28 | 北京波谱华光科技有限公司 | Video synthesis method, system and computer storage medium |
CN111586354A (en) * | 2020-04-28 | 2020-08-25 | 上海市保安服务(集团)有限公司 | Investigation system |
CN116320535A (en) * | 2023-04-14 | 2023-06-23 | 北京百度网讯科技有限公司 | Method, device, electronic equipment and storage medium for generating video |
CN116320535B (en) * | 2023-04-14 | 2024-03-22 | 北京百度网讯科技有限公司 | Method, device, electronic equipment and storage medium for generating video |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104239501B (en) | Mass video semantic annotation method based on Spark | |
US20180025020A1 (en) | System and method for contextually enriching a concept database | |
Xu et al. | The big data analytics and applications of the surveillance system using video structured description technology | |
CN110598042A (en) | Incremental update-based video structured real-time updating method and system | |
CN111090763B (en) | Picture automatic labeling method and device | |
CN111967302A (en) | Video tag generation method and device and electronic equipment | |
CN101853377A (en) | Method for identifying content of digital video | |
Karthikeyan et al. | Probability based document clustering and image clustering using content-based image retrieval | |
CN112257801A (en) | Incremental clustering method and device for images, electronic equipment and storage medium | |
Tian et al. | MCA-NN: Multiple correspondence analysis based neural network for disaster information detection | |
CN111882034A (en) | Neural network processing and face recognition method, device, equipment and storage medium | |
CN114491034B (en) | Text classification method and intelligent device | |
Ushapreethi et al. | Action recongnition in video survillance using hipi and map reducing model | |
CN113313098B (en) | Video processing method, device, system and storage medium | |
Gayathri et al. | Improved fuzzy-based SVM classification system using feature extraction for video indexing and retrieval | |
Negi et al. | Object detection based approach for an efficient video summarization with system statistics over cloud | |
CN115953650A (en) | Training method and device of feature fusion model | |
CN110083654A (en) | A kind of multi-source data fusion method and system towards science and techniques of defence field | |
Moses et al. | A classified study on semantic analysis of video summarization | |
Nagaraja et al. | Content based video retrieval using support vector machine classification | |
CN111178409B (en) | Image matching and recognition system based on big data matrix stability analysis | |
CN114049582A (en) | Weak supervision behavior detection method and device based on network structure search and background-action enhancement | |
CN111581420A (en) | Medical image real-time retrieval method based on Flink | |
CN107992474B (en) | Stream data theme mining method and system | |
WO2015197628A1 (en) | Method and device for analyzing a database comprising a large number of video images |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191220 |
|
RJ01 | Rejection of invention patent application after publication |