CN109313709A - A kind of measure of similarity, device and storage device - Google Patents
A kind of measure of similarity, device and storage device Download PDFInfo
- Publication number
- CN109313709A CN109313709A CN201780036577.XA CN201780036577A CN109313709A CN 109313709 A CN109313709 A CN 109313709A CN 201780036577 A CN201780036577 A CN 201780036577A CN 109313709 A CN109313709 A CN 109313709A
- Authority
- CN
- China
- Prior art keywords
- feature
- difference
- similarity
- default template
- treated
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
This application discloses a kind of measure of similarity, device and storage devices, are related to identification technology field.The described method includes: obtaining the feature of object to be identified；The difference between object feature and the feature of default template is calculated, and the difference is handled using preset strategy, difference is greater than or equal to the difference before processing so that treated；Using treated, difference calculates the similarity between object and default template.By the above-mentioned means, the application more acurrate quickly can carry out identification classification to object to be identified, recognition speed and discrimination are improved.
Description
Technical field
The present invention relates to identification technology fields, more particularly to a kind of measure of similarity, device and have storage
The device of function.
Background technique
When carrying out identifying processing to some information, usually by calculating certain features in information, these features pair
There may be certain discrimination for different targets, then these features are compared with the feature of default template again,
To complete the identification classification to information.For example, in some simple classification scenes, we only need given threshold value can be into
Row identification is distinguished.But present inventor is in long-term R&D process, finds the recognition speed of this method and accurate
Rate is lower, and given such as threshold value is likely to result in misclassification problem, especially in some relative complex scenes, it is given not
It is more similar with the feature between default template, it easily causes and obscures, not only easily cause identification mistake, cause misclassification, also
Recognition rate can be reduced, or even judgement not can be carried out to some features obscured that may cause.
Summary of the invention
The invention mainly solves the technical problem of providing a kind of measure of similarity, device and there is store function
Device, can be improved recognition speed and discrimination.
In order to solve the above technical problems, one technical scheme adopted by the invention is that: a kind of measurement side of similarity is provided
Method, which comprises obtain the feature of object to be identified；The difference between object feature and the feature of default template is calculated,
And the difference is handled using preset strategy, so that treated, difference is greater than or equal to the difference before processing；It utilizes
Difference that treated calculates the similarity between object and default template.
In order to solve the above technical problems, another technical solution used in the present invention is: providing a kind of measurement of similarity
Device, described device include processor, memory and telecommunication circuit, and processor couples memory and telecommunication circuit；Processor exists
When work, the feature of object to be identified is obtained by telecommunication circuit, is then calculated between object feature and the feature of default template
Difference, and the difference is handled using preset strategy, so that treated, difference is greater than or equal to the difference before processing
Value；Using treated, difference calculates the similarity between object and default template.
In order to solve the above technical problems, another technical solution that the application uses is: providing a kind of with store function
Device, described device is stored with program, and described program is performed the measure for realizing above-mentioned similarity.
The beneficial effects of the present invention are: being in contrast to the prior art, the application is to object to be identified and default mould
It when carrying out measuring similarity between plate, is handled by the difference between feature, making that treated, difference is greater than or equal to processing
Preceding difference, can difference between amplification characteristic, identification classification quickly is carried out to object to be identified so as to more acurrate,
Improve recognition speed and discrimination.
Detailed description of the invention
Fig. 1 is the flow diagram of the measure first embodiment of the application similarity；
Fig. 2 is the flow diagram of the measure second embodiment of the application similarity；
Fig. 3 is the template coding mode of one-dimension code in UPC-A code；
Fig. 4 is the structural schematic diagram of the measurement apparatus first embodiment of the application similarity；
Fig. 5 is the structural schematic diagram for the device first embodiment that the application has store function.
Specific embodiment
It is right as follows in conjunction with drawings and embodiments to keep the purpose, technical solution and effect of the application clearer, clear
The application is further described.
The application provides the measure and device of a kind of similarity, may at least apply for image recognition processing scene
In, feature is more similar especially between given multiple default templates holds in confusing scene.By calculating wait know
When similarity between other image and default template, the difference feature is handled, with the difference between amplification characteristic, thus
More acurrate identification classification quickly can be carried out to images to be recognized, improve recognition speed and discrimination.Lower mask body expansion is said
It is bright:
Referring to Fig. 1, Fig. 1 is the flow diagram of the measure first embodiment of the application similarity.Such as Fig. 1 institute
Show, in this embodiment, the measure of similarity includes:
S101: the feature of object to be identified is obtained.
In this step, object to be identified can be image, such as one-dimension code image, image in 2 D code.Acquired object
Feature can be at least one of area, width, perimeter, density, these object features have for different default templates
There is certain discrimination, by the comparison to these object features and default template characteristic, which can be carried out
Identification classification.
S102: the difference between object feature and the feature of default template is calculated, and using preset strategy to the difference
It is handled, so that treated, difference is greater than or equal to the difference before processing.
Specifically, default template corresponding with object to be identified is selected, object to be identified and default template are subjected to phase
Compared like degree, by calculating the difference between object feature and default template characteristic, calculate object to be identified and default template it
Between similarity.Wherein, when calculating the difference between object feature and the feature of default template, using preset strategy to described
Difference is handled, so that treated, difference is greater than or equal to the difference before processing.By processing difference, difference is amplified,
The difference between feature can be expanded, be equivalent to and a punishment has been carried out to original similarity；Object feature can be reduced and preset
Similarity between template characteristic reaches preferable global similitude.It, can be to all features when handling difference
Difference is handled, and can also only be handled the difference of Partial Feature.
Wherein, the smaller explanation object feature of difference and default template characteristic between object feature and default template characteristic
More close, then the similarity between them is bigger, and corresponding measuring similarity value is also bigger；On the contrary, object feature and pre-
If bigger explanation object feature of difference between template characteristic is more become estranged with default template characteristic, then similar between them
Degree is just smaller, and corresponding measuring similarity value is also smaller.Wherein, the difference between object feature and default template characteristic can be with
It is calculated using similarity factor function and distance function etc..
For example, calculating in object to be identified feature A1 at a distance from feature A in default template according to conventional distance function
Difference is 3, in the measure of the similarity of the application, double processing can be carried out to the difference, by the flat of normal difference
Fang Zuowei last difference, then last difference just becomes 9 (3 squares), after dealing in this way, object feature and default mould
Difference between plate features becomes larger, also just more dissimilar, it is easier to distinguish similar similar feature.Improve recognition speed and knowledge
Not rate.
S103: using treated, difference calculates the similarity between object and default template.
After processing obtains the difference between feature, the difference of comprehensive all features is calculated between object to be identified and default template
Similarity.
Referring to Fig. 2, Fig. 2 is the flow diagram of the measure second embodiment of the application similarity.In the reality
It applies in mode, after the feature for obtaining object to be identified, first these object features is clustered, carry out similarity after cluster again
Measurement.By first clustering to feature, calculating step can be simplified, improve recognition speed and discrimination.As shown in Fig. 2,
In this embodiment, the measure of similarity includes:
S201: the feature of object to be identified is obtained.
S202: the metric of object feature to be identified is calculated, and the object feature is clustered.
The metric for calculating each object feature (such as width) first can use Binarization methods, Gabor wavelet becomes
Scaling method or depth convolutional network algorithm etc. calculate the metric of each object feature.These characteristic values can be combined into n tie up to
Amount, the number that wherein n is characterized.
After obtaining the metric of each object feature, the object feature is clustered.It can according to application environment-identification
Feature is divided into two classes, four classes etc..K-menas clustering algorithm, Otsu algorithm (OSTU) or density algorithm be can use to visitor
Body characteristics are clustered.
S203: the difference between object feature and the feature of default template is calculated, and using preset strategy to the difference
It is handled, so that treated, difference is greater than or equal to the difference before processing.
After obtaining the classification results of feature, default template corresponding with object to be identified is selected to carry out similarity ratio
It is right, calculate the similarity between object and default template.For different identification scenes, calculate between object and default template
The mode of similarity is different, can be simple distance calculating or sequence etc. for simple application scenarios, also can use phase
The similarity between object to be identified and default template is calculated like coefficient function or distance function.
Wherein, when calculating the measuring similarity value between feature, such as when calculating distance difference, using preset strategy to away from
Deviation value is handled, and preset strategy can be the square value for calculating distance difference, cube value；Distance difference is added or is passed through
Other formula algorithms amplify the distance difference；So that treated, distance difference is greater than or equal to the distance difference before processing, puts
Difference between big feature, reduces the similarity between feature.It in other embodiments, can also be by last measuring similarity value
After disposed of in its entirety is carried out to get overall similarity metric is arrived, square value, cube value of overall similarity metric etc. are calculated.
Optionally, formula is utilizedCalculate object feature to be identified and default mould
The distance between plate features difference, wherein d is treated total distance difference, the number that M is characterized, riFor in default template
The characteristic value of ith feature, wiFor the characteristic value of ith feature in object to be identified.Wherein, which is suitable for object feature
The case where the distance between default template characteristic difference is greater than 1, when distance difference is greater than 1, by distance difference calculating square
Afterwards, the distance difference can be made to increase, so that the difference between feature is become larger, similarity becomes smaller.
S204: using treated, difference calculates the similarity between object and default template.
S205: the similarity between object and default template is carried out similar between sequencing of similarity or exclusion and object
Degree is less than the default template of preset threshold.
After the similarity being calculated between object and default template, gained similarity can be handled, it is such as right
Similarity sorts from high to low, or excludes the default template etc. that similarity is less than preset threshold, wherein can be answered according to different
With scene, the difference size between template is preset, the preset threshold is arranged in adaptability.As similar between object and default template
Spend higher, then illustrating that the object object has very big may be to belong to same class with default template；And object and default template
Between similarity it is lower, then illustrate the object object may be not belonging to default template that is a kind of.In this way, can
Discrimination is greatly improved, the object relatively unambiguous for feature can directly obtain recognition result.
In an application scenarios, the measure of similarity provided herein can be applied to identification one-dimension code figure
Picture.Specifically,
One-dimension code is usually made of the not equal black and white item of some width, for simple commodity code, such as UPC-A code, it
Each character be to be made of two secret notes and two informal vouchers；The width of the smallest secret note of width or informal voucher is called module,
The overall width of so one character is 7 modules.Allowing the width of secret note and informal voucher is respectively 1,2,3,4 times of a module,
So a character is made of four width, and each width means are several times of module width, the width of different character black and white items
Degree combination is different.UPC-A code only supports 0-9 to have 10 numbers altogether, and each number has different width coding modes, please refers to
Fig. 3, Fig. 3 are the width coding modes of one-dimension code in UPC-A code, as shown in figure 3, the width coding mode of number 0-9 is respectively,
Digital 0:(3,2,1,1)；Digital 1:(2,2,2,1)；Digital 2:(2,1,2,2)；Digital 3:(1,4,1,1)；Digital 4:(1,1,3,
2)；Digital 5:(1,2,3,1)；Digital 6:(1,1,1,4)；Digital 7:(1,3,1,2)；Digital 8:(1,2,1,3)；Digital 9:(3,
1,1,2)。
For bar code image secondary for one, simple feature, i.e. width are calculated first.The width of each black and white item can be two
It is counted after value.Statistical method is can be obtained by by simple statistical pixel number.
After obtaining width value, can classify to width, such as above-mentioned bar code include 4 kinds of width, then then to width into
Row classification, can be used any classification method, such as simply be clustered using kmenas, each black and white that will be indicated with pixel
The width of item is divided into 1,2,3,4 4 classes.
After tagsort, the code width of these features and one-dimension code is compared into calculating similarity.It utilizes formula (1)
Punishment processing is carried out to the similarity of k-th of character width and code width:
Wherein, d is treated total difference,Width for coding k i-th of position,It is the i-th of k-th of character
Root black and white width, M are the number of width.Difference is bigger, and the difference between feature is bigger, then similarity is lower.The formula is applicable in
The case where width difference in some position is greater than 1, because calculate width difference square can make the width when width difference is greater than 1
Difference becomes larger.For example the width requirement of i-th black and white item of k-th of character is 2 in code width, and calculate the sorted knot of width
Fruit is 4, then the not simple classification deviation that calculates is 2 (4 subtract 2), and punishment is 4 (2 multiplied by 2), that is to say, that thinks width
In the case where differing by more than 1, a possibility that not being the coding, is bigger.
For example, the width classification results for obtaining a character are 2,2,2,1, then the width similarity with code character 1
Punishment is 0, especially character 3 then larger with the punishment of other templates, 6,8, comprising with code width difference incessantly for 1 it is black
Item or informal voucher.It thus can learn the character representation number 0, or the very low coding templet of similarity can be excluded (if not being
The possibility of number 3 is very big).
After obtaining the similarity between character and coding templet, these similarities can be handled, such as to similarity
Sort from high to low, exclude the very low coding templet etc. of similarity, as the similarity of this character and 1 or 2 coding templet compared with
Height, then the character may be that the probability of number 1 or 2 is larger, and the similarity of this character and 3 or 6 coding templet is lower, then
The character is that the probability of number 3 or 6 is smaller, can quickly identify character or exclude dissimilar coding, improve recognition speed and
Discrimination.For UPC-A, only 10 kinds of coding modes altogether, and for code128, different width coding templates
Can be up to a hundred, dissimilar coding mode can be excluded significantly using this method, so that discrimination is greatly improved, for bar code matter
Higher image is measured, recognition result can be directly obtained.
Referring to Fig. 4, Fig. 4 is the structural schematic diagram of the measurement apparatus first embodiment of the application similarity.The present embodiment
In the measurement apparatus of similarity the measure of above-mentioned similarity may be implemented, which includes processor 401, memory
402 and telecommunication circuit 403.Processor 401 couples memory 402 and telecommunication circuit 403, and processor 401 executes at work to be referred to
It enables, to cooperate memory 402 and telecommunication circuit 403 to realize the measure of above-mentioned similarity, specific work process and above-mentioned side
It is consistent in method embodiment, therefore details are not described herein, please refers to the explanation of the above corresponding method step in detail.Wherein, similarity
Measurement apparatus can be screen bar code recognizer, image analyzer etc..
Referring to Fig. 5, Fig. 5 is the structural schematic diagram for the device first embodiment that the application has store function.This reality
It applies storage device 50 in example and is stored with program 501, program 501 is performed the measure for realizing above-mentioned similarity.Specific work
Make in process and above method embodiment unanimously, therefore details are not described herein, please refers to the explanation of the above corresponding method step in detail.
Wherein with store function device can be portable storage media for example USB flash disk, CD, mobile hard disk, read-only memory (ROM,
Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk etc. is various can store
The medium of program code is also possible to terminal, server etc..
Above scheme, the application pass through feature when to measuring similarity is carried out between object to be identified and default template
Between difference handled, making that treated, difference is greater than or equal to the difference before processing, can difference between amplification characteristic,
Identification classification quickly is carried out to object to be identified so as to more acurrate, improves recognition speed and discrimination.
In several embodiments provided herein, it should be understood that disclosed system, device and method can
To realize by another way.For example, device embodiments described above are only schematical, for example, the mould
The division of block or unit, only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple
Unit or assembly can be combined or can be integrated into another system, or some features can be ignored or not executed.It is another
Point, shown or discussed mutual coupling, direct-coupling or communication connection can be through some interfaces, device or
The indirect coupling or communication connection of unit can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.Some or all of unit therein can be selected to realize present embodiment scheme according to the actual needs
Purpose.
In addition, each functional unit in each embodiment of the application can integrate in one processing unit, it can also
To be that each unit physically exists alone, can also be integrated in one unit with two or more units.It is above-mentioned integrated
Unit both can take the form of hardware realization, can also realize in the form of software functional units.
If the integrated unit is realized in the form of SFU software functional unit and sells or use as independent product
When, it can store in a computer readable storage medium.Based on this understanding, the technical solution of the application is substantially
The all or part of the part that contributes to existing technology or the technical solution can be in the form of software products in other words
It embodies, which is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the application
The all or part of the steps of embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory
(ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic or disk
Etc. the various media that can store program code.
The foregoing is merely presently filed embodiments, are not intended to limit the scope of the patents of the application, all to utilize this
Equivalent structure or equivalent flow shift made by application specification and accompanying drawing content, it is relevant to be applied directly or indirectly in other
Technical field is included within the scope of the present invention.
Claims (19)
1. a kind of measure of similarity, which is characterized in that the described method includes:
Obtain the feature of object to be identified；
Calculate the difference between the object feature and the feature of default template, and using preset strategy to the difference at
Reason, so that treated, difference is greater than or equal to the difference before processing；
Using treated, difference calculates the similarity between the object and the default template.
2. the method according to claim 1, wherein between calculating object feature and the feature of default template
Difference, the difference is handled using preset strategy, so that treated difference is greater than or equal to the difference before processing
Include:
Utilize formulaIt calculates between the object feature and the default template characteristic
Difference, wherein d is treated total difference, the number that M is characterized, riFor the characteristic value for presetting ith feature in template, wi
For the characteristic value of ith feature in object to be identified, and (ri-wi) it is greater than 1.
3. the method according to claim 1, wherein between calculating object feature and the feature of default template
Difference before include: to calculate the metric of the object feature, and cluster to the object feature.
4. according to the method described in claim 3, it is characterized in that, between calculating object feature and the feature of default template
Difference include: using distance function calculate cluster after the object feature and the distance between the feature of default template it is poor
Value.
5. according to the method described in claim 3, it is characterized in that, it is described to object feature carry out cluster include: to utilize k-
Menas clustering algorithm, Otsu algorithm or density algorithm cluster the object feature.
6. according to the method described in claim 3, it is characterized in that, the metric for calculating object feature includes: to utilize two
Value algorithm, Gabor wavelet transformation algorithm or depth convolutional network calculate the metric of the object feature.
7. the method according to claim 1, wherein described, using treated, difference calculates the object and institute
State the similarity between default template includes: later
Similarity between the object and the default template is carried out between sequencing of similarity or exclusion and the object
Similarity is less than the default template of preset threshold.
8. the method according to claim 1, wherein the object feature is area, in width, perimeter, density
At least one.
9. the method according to claim 1, wherein the object to be identified is one-dimension code image.
10. a kind of measurement apparatus of similarity, which is characterized in that described device includes processor, memory and telecommunication circuit, institute
It states processor and couples the memory and telecommunication circuit；
The processor at work, the feature of object to be identified is obtained by the telecommunication circuit, then calculates the object
Difference between feature and the feature of default template, and the difference is handled using preset strategy, so that treated
Difference is greater than or equal to the difference before processing；Using treated, difference calculates the phase between the object and the default template
Like degree.
11. device according to claim 10, which is characterized in that the feature for calculating object feature and default template it
Between difference, the difference is handled using preset strategy, so that treated difference is greater than or equal to the difference before processing
Value includes:
The processor at work, utilizes formulaCalculate the object feature and institute
State the difference between default template characteristic, wherein d is treated total difference, the number that M is characterized, riFor in default template
The characteristic value of ith feature, wiFor the characteristic value of ith feature in object to be identified, and (ri-wi) it is greater than 1.
12. device according to claim 10, which is characterized in that the feature for calculating object feature and default template it
Between difference before include:
The processor at work, calculates the metric of the object feature, and cluster to the object feature.
13. device according to claim 12, which is characterized in that the feature for calculating object feature and default template it
Between difference include:
The processor at work, the spy of the object feature and the default template after clustering is calculated using distance function
The distance between sign difference.
14. device according to claim 12, which is characterized in that it is described to object feature carry out cluster include: to utilize k-
Menas clustering algorithm, Otsu algorithm or density algorithm cluster the object feature.
15. device according to claim 12, which is characterized in that the metric for calculating object feature includes: to utilize
Binarization methods, Gabor wavelet transformation algorithm or depth convolutional network calculate the metric of the object feature.
16. device according to claim 10, which is characterized in that it is described using treated difference calculate the object with
Include: after similarity between the default template
The processor carries out sequencing of similarity or row at work, to the similarity between the object and the default template
Except the similarity between the object is less than the default template of preset threshold.
17. device according to claim 10, which is characterized in that the object feature is area, width, perimeter, density
At least one of.
18. device according to claim 10, which is characterized in that the object to be identified is one-dimension code image.
19. a kind of device with store function, which is characterized in that described device is stored with program, and described program is performed
Realize the measure of the described in any item similarities of claim 1 to 9.
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2017/120231 WO2019127504A1 (en) | 2017-12-29 | 2017-12-29 | Similarity measurement method and device, and storage device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109313709A true CN109313709A (en) | 2019-02-05 |
Family
ID=65225809
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201780036577.XA Pending CN109313709A (en) | 2017-12-29 | 2017-12-29 | A kind of measure of similarity, device and storage device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109313709A (en) |
| WO (1) | WO2019127504A1 (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117891883B (en) * | 2024-03-14 | 2024-06-18 | 山东观和集团有限公司 | Mineral exploration data optimal storage method |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408932A (en) * | 2008-04-11 | 2009-04-15 | 浙江师范大学 | Method for matching finger print image based on finger print structure feature and veins analysis |
| CN101604384A (en) * | 2009-07-17 | 2009-12-16 | 周易 | Individualized fingerprint identification method |
| US20110185234A1 (en) * | 2010-01-28 | 2011-07-28 | Ira Cohen | System event logs |
| CN103049750A (en) * | 2013-01-11 | 2013-04-17 | 广州广电运通金融电子股份有限公司 | Character recognition method |
| CN103336945A (en) * | 2013-06-10 | 2013-10-02 | 黑龙江大学 | Finger vein recognition method fusing local features and global features |
| CN104143086A (en) * | 2014-07-18 | 2014-11-12 | 吴建忠 | Application technology of portrait comparison to mobile terminal operating system |
| CN104766343A (en) * | 2015-03-27 | 2015-07-08 | 电子科技大学 | Vision target tracking method based on sparse representation |
| CN104951940A (en) * | 2015-06-05 | 2015-09-30 | 西安理工大学 | Mobile payment verification method based on palmprint recognition |
| CN105787451A (en) * | 2016-02-29 | 2016-07-20 | 南京邮电大学 | Fingerprint matching method based on multi-judgment point mode |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426176B (en) * | 2013-08-27 | 2017-03-01 | 重庆邮电大学 | Based on the shot detection method improving rectangular histogram and clustering algorithm |
| CN106651777B (en) * | 2015-10-29 | 2020-07-03 | 小米科技有限责任公司 | Image processing method and device and electronic equipment |
| CN107203686B (en) * | 2017-03-31 | 2021-04-20 | 苏州艾隆信息技术有限公司 | Medicine information difference processing method and system |
-
2017
- 2017-12-29 CN CN201780036577.XA patent/CN109313709A/en active Pending
- 2017-12-29 WO PCT/CN2017/120231 patent/WO2019127504A1/en active Application Filing
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101408932A (en) * | 2008-04-11 | 2009-04-15 | 浙江师范大学 | Method for matching finger print image based on finger print structure feature and veins analysis |
| CN101604384A (en) * | 2009-07-17 | 2009-12-16 | 周易 | Individualized fingerprint identification method |
| US20110185234A1 (en) * | 2010-01-28 | 2011-07-28 | Ira Cohen | System event logs |
| CN103049750A (en) * | 2013-01-11 | 2013-04-17 | 广州广电运通金融电子股份有限公司 | Character recognition method |
| CN103336945A (en) * | 2013-06-10 | 2013-10-02 | 黑龙江大学 | Finger vein recognition method fusing local features and global features |
| CN104143086A (en) * | 2014-07-18 | 2014-11-12 | 吴建忠 | Application technology of portrait comparison to mobile terminal operating system |
| CN104766343A (en) * | 2015-03-27 | 2015-07-08 | 电子科技大学 | Vision target tracking method based on sparse representation |
| CN104951940A (en) * | 2015-06-05 | 2015-09-30 | 西安理工大学 | Mobile payment verification method based on palmprint recognition |
| CN105787451A (en) * | 2016-02-29 | 2016-07-20 | 南京邮电大学 | Fingerprint matching method based on multi-judgment point mode |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019127504A1 (en) | 2019-07-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109829398B (en) | Target detection method in video based on three-dimensional convolution network | |
| Odone et al. | Building kernels from binary strings for image matching | |
| US8873856B1 (en) | Determining a class associated with an image | |
| CN106228129B (en) | A kind of human face in-vivo detection method based on MATV feature | |
| CN111832650B (en) | Image classification method based on generation of antagonism network local aggregation coding semi-supervision | |
| Ito et al. | Object classification using heterogeneous co-occurrence features | |
| CN108875727B (en) | The detection method and device of graph-text identification, storage medium, processor | |
| JP6401648B2 (en) | Defect classification apparatus and defect classification method | |
| CN110223310B (en) | Line structure light center line and box edge detection method based on deep learning | |
| CN109583493A (en) | A kind of credit card detection and digit recognition method based on deep learning | |
| CN109766942A (en) | A kind of small-sample learning image-recognizing method based on attention neural network | |
| US20210327041A1 (en) | Image based novelty detection of material samples | |
| CN106295710B (en) | Image local feature matching process, device and terminal based on non-geometric constraint | |
| CN112633392A (en) | Terahertz human body security inspection image target detection model training data augmentation method | |
| CN110991201B (en) | Bar code detection method and related device | |
| CN111444816A (en) | Multi-scale dense pedestrian detection method based on fast RCNN | |
| CN112241736A (en) | Text detection method and device | |
| CN112070116B (en) | Automatic artistic drawing classification system and method based on support vector machine | |
| CN109313709A (en) | A kind of measure of similarity, device and storage device | |
| Yang et al. | License plate detection based on sparse auto-encoder | |
| Gad et al. | Crowd density estimation using multiple features categories and multiple regression models | |
| Khuspe et al. | Robust image forgery localization and recognition in copy-move using bag of features and SVM | |
| JP6151908B2 (en) | Learning device, identification device, and program thereof | |
| Sanin et al. | K-tangent spaces on Riemannian manifolds for improved pedestrian detection | |
| Khairuddin et al. | Using two stage classification for improved tropical wood species recognition system |
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 | ||
| CB02 | Change of applicant information |
Address after: 518063 23 Floor (Room 2303-2306) of Desai Science and Technology Building, Yuehai Street High-tech Zone, Nanshan District, Shenzhen City, Guangdong Province Applicant after: SHENZHEN A&E INTELLIGENT TECHNOLOGY INSTITUTE Co.,Ltd. Address before: 518104 Shajing Industrial Co., Ltd. No. 3 Industrial Zone, Hexiang Road, Shajing Street, Baoan District, Shenzhen City, Guangdong Province Applicant before: SHENZHEN A&E INTELLIGENT TECHNOLOGY INSTITUTE Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190205 |
|
| RJ01 | Rejection of invention patent application after publication |