CN104077594A - Image recognition method and device - Google Patents
Image recognition method and device Download PDFInfo
- Publication number
- CN104077594A CN104077594A CN201310109749.5A CN201310109749A CN104077594A CN 104077594 A CN104077594 A CN 104077594A CN 201310109749 A CN201310109749 A CN 201310109749A CN 104077594 A CN104077594 A CN 104077594A
- Authority
- CN
- China
- Prior art keywords
- image
- identified
- face
- rule
- original image
- 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.)
- Granted
Links
Abstract
The invention discloses an image recognition method and device and aims at solving the problem of high false detection rate existing in the prior art during image recognition. The image recognition method comprises the steps of dividing images to be recognized to obtain multiple image blocks according to a preset division rule, wherein the division rule meet the requirement that the obtained image blocks include pixel points capable of representing characteristics of targets to be recognized in the images to be recognized; respectively performing characteristic extraction on multiple image blocks to obtain characteristic data of the image blocks; classifying the image blocks to obtain classification results of the image blocks by using a classifier obtained according to the characteristic data of the obtained image blocks and the preset division rule and used for distinguishing image block categories; determining the block categories according to the classification results of the image blocks and a preset determination rule.
Description
Technical field
The present invention relates to area of pattern recognition, relate in particular to a kind of image-recognizing method and device.
Background technology
At present, mode identification technology has obtained applying more and more widely, and it has a wide range of applications at numerous areas such as safety, finance, man-machine interaction, information, education.Existing mode identification technology, when carrying out image recognition, generally will be chosen suitable feature according to image itself, just can obtain good recognition effect.For example, in face recognition process, generally by whole image input face identification device to be identified, and the sample image in the image pattern storehouse of the image of input and face identification device is compared, if the characteristic of this whole image is not mated with the characteristic of sample image in image pattern storehouse, the image of this input is defined as to abnormal image, and the people's face in the image of this input is defined as to abnormal face.Wherein, abnormal face generally refers to has worn people's face that can shelter from face organ's accessories, and normal person's face is for abnormal face, and all non-abnormal faces are all considered as normal person's face.
There is following defect in above-mentioned existing method: false drop rate is higher.Such as, when the image of people's face of identification wear dark glasses, the latter half of people's face in image is the same with normal person's face, this just causes the feature of the feature of whole image and the image of normal person's face more approaching, and the feature of the image of wear dark glasses part will more not given prominence to, thereby will be easy to cause flase drop, this abnormal image is judged as to normal picture.
Summary of the invention
The embodiment of the present invention provides a kind of image-recognizing method and device, in order to solve the higher problem of false drop rate existing in prior art when image is identified.
The embodiment of the present invention is by the following technical solutions:
An image-recognizing method, comprising:
According to predefined division rule, treat recognition image and divide, obtain a plurality of image blocks; Wherein, described division rule meets: in each image block that makes to obtain, all comprise the pixel that can characterize the clarification of objective to be identified in described image to be identified; And
Described a plurality of image blocks are carried out respectively to feature extraction, obtain the characteristic of each image block;
According to the characteristic of each image block obtaining, and according to described division rule, train the sorter for differentiate between images piece classification obtaining in advance, respectively each image block is classified, obtain the classification results of each image block;
According to the classification results of each image block and default decision rule, determine the classification of described image.
A pattern recognition device, comprising:
Division unit, for according to predefined division rule, treats recognition image and divides, and obtains a plurality of image blocks; Wherein, described division rule meets: in each image block that makes to obtain, all comprise the pixel that can characterize the clarification of objective to be identified in described image to be identified;
Feature extraction unit, carries out respectively feature extraction for described a plurality of image blocks that division unit is obtained, and obtains the characteristic of each image block;
Taxon, characteristic for each image block of obtaining according to feature extraction unit, and according to described division rule, train the sorter for differentiate between images piece classification obtaining in advance, and respectively each image block is classified, obtain the classification results of each image block;
Classification determining unit, for the classification results of each image block of obtaining according to taxon and default decision rule, determines the classification of described image.
The beneficial effect of the embodiment of the present invention is as follows:
The embodiment of the present invention is identified respectively and is classified by each image block obtaining after whole image divided, when image block under the unusual part in abnormal image is identified, can not be subject to the impact of the affiliated image block of other normal parts, thereby the classification results of each image block is all more accurate, thereby the accuracy rate of the recognition result of whole image finally obtaining also can be improved, avoided in prior art due to the feature of the image of the unusual part feature outstanding problem that causes flase drop with respect to whole image.
Accompanying drawing explanation
The main process flow diagram of a kind of image-recognizing method that Fig. 1 provides for the embodiment of the present invention one;
The particular flow sheet of a kind of face identification method that Fig. 2 provides for the embodiment of the present invention two;
The division schematic diagram of the image to be identified that Fig. 3 provides for the embodiment of the present invention two;
The structural representation of a kind of pattern recognition device that Fig. 4 provides for the embodiment of the present invention three.
Embodiment
In order to solve the higher problem of false drop rate existing in prior art when image is identified, the embodiment of the present invention provides a kind of image recognition scheme.This scheme is identified respectively and is classified by each image block obtaining after whole image divided, when image block under the unusual part in abnormal image is identified, can not be subject to the impact of the affiliated image block of other normal parts, thereby the classification results of each image block is all more accurate, thereby the accuracy rate of the recognition result of whole image finally obtaining also can be improved, also avoided in prior art due to the feature of the image of the unusual part feature outstanding problem that causes flase drop with respect to whole image simultaneously.
Below in conjunction with each accompanying drawing, embodiment of the present invention technical scheme main realized to principle, embodiment and the beneficial effect that should be able to reach is explained in detail.
Embodiment mono-:
As shown in Figure 1, the main process flow diagram of a kind of image-recognizing method providing for the embodiment of the present invention, the method comprises the following steps:
Step 11, according to predefined division rule, treats recognition image and divides, and obtains a plurality of image blocks;
Concrete, under different application scenarios and image type, can there is different division methods to treat recognition image and divide.Such as, when facial image is identified, can be divided into two image blocks by along continuous straight runs, be respectively first face image block and second face image block; Or, also can along continuous straight runs facial image be divided into three or four image blocks.When other images are identified, also can vertically treat recognition image divides, or even divide along the oblique recognition image for the treatment of of miter angle, this is not restricted for the concrete dividing mode embodiment of the present invention, can set voluntarily according to actual demand.
Wherein, division rule should meet: in each image block that makes to obtain, all comprise the pixel that can characterize the clarification of objective to be identified in image to be identified.
In addition, image to be identified can be directly to obtain, and also can determine by original image.Wherein, when the image receiving is original image, a kind of mode of definite image to be identified can specifically comprise the steps:
First, the edge pixel of the target to be identified in the original image receiving by detection point, determines the area of target to be identified;
Whether the area ratio that then, judges target to be identified and original image is greater than default proportion threshold value;
When judgment result is that while being, original image is defined as to image to be identified;
When the determination result is NO, according to default rectangular area, determine rule, from original image, determine the rectangular area that comprises target to be identified, and rectangular area is defined as to image to be identified, wherein, rectangular area determines that rule should meet: make target to be identified and the area ratio of the rectangular area of determining be greater than above-mentioned proportion threshold value.
When target behaviour face to be identified, can determine image to be identified according to above-mentioned definite mode, also can determine according to following definite mode, specifically comprise:
First, receive original image, and detect in this original image, whether to comprise people's face;
Then, while comprising people's face in detecting original image, from original image, determine the extraneous rectangle region of people's face;
Finally, from original image, be partitioned into above-mentioned zone, and be defined as image to be identified.
It should be noted that, above-mentioned mention according to rectangular area, determine the rectangular area that rule is determined, and the boundary rectangle region of the people's face being partitioned into from original image, can be first scaled to default big or small image, then the image that carries out obtaining after convergent-divergent is defined as to image to be identified.
Step 12, carries out respectively feature extraction to a plurality of image blocks that obtain, and obtains the characteristic of each image block;
Wherein, can be according to edge orientation histogram, local binary (Local Binary Pattern, LBP), yardstick invariant features conversion (Scale Invariant Feature Transform, SIFT) and accelerate the characteristic that the features such as robust features (Speeded-Up Robust Feature, SURF) obtain each image block.
Step 13, according to the characteristic of each image block obtaining, and trains according to division rule the sorter for differentiate between images piece classification obtaining in advance, respectively each image block is classified, and obtains the classification results of each image block.
Step 14, according to the classification results of each image block and default decision rule, determines the classification of this image.
The embodiment of the present invention is identified respectively and is classified by each image block obtaining after whole image divided, when image block under the unusual part in abnormal image is identified, can not be subject to the impact of the affiliated image block of other normal parts, thereby the classification results of each image block is all more accurate, thereby the accuracy rate of the recognition result of whole image finally obtaining also can be improved, also avoided in prior art due to the feature of the image of the unusual part feature outstanding problem that causes flase drop with respect to whole image simultaneously.
Embodiment bis-:
The embodiment of the present invention two is elaborated to image-recognizing method in embodiment mono-in connection with practical application, in the embodiment of the present invention, take and above-mentioned image-recognizing method is applied to recognition of face describes as example, concrete application scenario can be the occasions such as ATM monitoring video automatic alarm.When having people when withdrawing cash, can carry out automatic classification to people's face, then according to its whether abnormal selection whether report to the police, and can link with ATM, when identifying people's face and be abnormal face, ATM is not told paper money.
The particular flow sheet of a kind of face identification method providing for the embodiment of the present invention as shown in Figure 2.
Step 21, receives original image;
Step 22, detects in original image whether comprise people's face; While comprising people's face in detecting original image, execution step 23, otherwise directly this original image is defined as to abnormal image.
Step 23 is determined the boundary rectangle of people's face, and this boundary rectangle region is split as image to be identified from original image from original image;
Wherein, determining also of above-mentioned image to be identified can be determined according to the another kind of mode of introducing in embodiment mono-step 11.
Step 24, according to predefined division rule, along continuous straight runs is divided the image to be identified of determining, obtains first face image block and second face image block;
Wherein, division rule should meet: the pixel that makes all to comprise in first face image block of obtaining after dividing and second face image block the feature of people's face that can characterize in image to be identified.
Such as the height of the image to be identified of determining with width is respectively H and W(is a high H pixel, a wide W pixel), can divide with division limits YO=H/2, this image to be identified is placed in to coordinate axis, as shown in Figure 3, x direction of principal axis represents width, y direction of principal axis represent height, by x axle from 0 to W-1, y axle from 0 to YO-1 part as second face image block, by x axle, from 0 to W-1, the part of y axle from YO to H-1 is as first face image block.
Step 25, carries out respectively feature extraction to first face image block and second face image block, obtains the characteristic of first face image block and the characteristic of second face image block;
Wherein, for first face image block, to extract the example that is characterized as of edge orientation histogram, detailed process is as follows:
First face image block obtaining is carried out to rim detection, obtain edge intensity value computing and the edge direction values of each pixel.
Such as using this edge detection operator of Sobel operator to carry out convolution operation to first face image block, thereby obtain x axle and the axial Sobel rim value of y.Wherein, the Sobel operator for x axle and y axle can be successively:
The edge intensity value computing of each pixel in this first face image block is the absolute value sum of the absolute value of Sobel rim value and the Sobel rim value of y axle of x axle.
The computing method of the edge direction values of each pixel in this first face image block are as follows:
Wherein, in formula (1), the edge direction values that Dir is pixel, YSobel is the Sobel rim value of y axle, XSobel is the Sobel rim value of x axle.
Wherein, the size of the edge intensity value computing of pixel determines whether this pixel is edge pixel point, if its edge intensity value computing is too small, this pixel is not just edge pixel point, while carrying out edge orientation histogram statistics, just can not add up.Therefore, the pixel that in the embodiment of the present invention, edge intensity level is greater than default pixel edge intensity value computing (hereinafter to be referred as predetermined threshold value) carries out the statistics of edge orientation histogram, the edge intensity value computing that is greater than each pixel of predetermined threshold value is mapped as to the numerical value in edge orientation histogram, and the computing formula of employing is as follows:
Wherein, in formula (2), Grad is the edge intensity value computing of pixel in image block, and SobelThres is default pixel edge intensity value computing, and Hist (Dir) is the numerical value in edge orientation histogram.
Finally, in first face image block obtaining according to statistics, the edge orientation histogram of pixel, determines its characteristic.
Same, consistent with said process for the feature extraction of second face image block, therefore not to repeat here.
It should be noted that, the method for feature extraction has a lot, considers that the embodiment of the present invention is that people's face is identified.And generally, when people's face is identified, when people's face tilts to some extent or turn to a little (being exactly face little on one side), be just easy to cause flase drop.And the problem producing when the method for this feature of edge orientation histogram just can avoid above-mentioned people's face to tilt is extracted in employing.Because when people's face tilts to some extent, the edge orientation histogram of the edge orientation histogram extracting and normal person's face is consistent.Therefore, preferably extract the method for edge orientation histogram in the embodiment of the present invention, can certainly be not limited to said method and adopt other feature extracting method, concrete can select voluntarily according to actual conditions.
Step 26, according to the characteristic of each image block obtaining, and in advance according to division rule train obtain for distinguishing the sorter of normal picture piece and abnormal image piece, respectively each image block is classified, obtain the classification results of each image block.
Wherein, the concrete grammar of training classifier can be: the image pattern of choosing some (two or three hundred), then the image pattern of choosing is divided respectively and feature extraction, obtain respectively the characteristic of some first face image blocks and the characteristic of second face image block, using it as two classifications, train respectively sorter and a sorter for second face image block for first face image block.About the concrete training method of sorter, have a lot, can select according to actual conditions, in the embodiment of the present invention, be not described in detail in this.
Step 27, according to the classification results of first face image block and second face image block, whether judgement wherein comprises abnormal image piece, if comprise, performs step 28, otherwise execution step 29.
Step 28, is defined as abnormal image by image to be identified.
Step 29, is defined as normal picture by image to be identified.
In the embodiment of the present invention, by each image block obtaining after whole image divided, identify respectively and classify, when image block under the unusual part in abnormal image is identified, can not be subject to the impact of the affiliated image block of other normal parts, thereby the accuracy rate of the classification results of each image block can improve greatly, and the accuracy rate of the recognition result of whole the image finally obtaining also can improve, also avoided in prior art due to the feature of the image of the unusual part feature outstanding problem that causes flase drop with respect to whole image simultaneously.
Embodiment tri-:
The embodiment of the present invention three also provides a kind of pattern recognition device, and the structural representation of this device as shown in Figure 4, comprising:
Division unit 41, for according to predefined division rule, treats recognition image and divides, and obtains a plurality of image blocks; Wherein, division rule meets: in each image block that makes to obtain, all comprise the pixel that can characterize the clarification of objective to be identified in image to be identified;
Feature extraction unit 42, carries out respectively feature extraction for a plurality of image blocks that division unit 41 is obtained, and obtains the characteristic of each image block;
Taxon 43, characteristic for each image block of obtaining according to feature extraction unit 42, and according to division rule, train the sorter for differentiate between images piece classification obtaining in advance, and respectively each image block is classified, obtain the classification results of each image block;
Classification determining unit 44, for the classification results of each image block of obtaining according to taxon 43 and default decision rule, determines the classification of this image.
Wherein, during target behaviour face to be identified in image to be identified, division unit 41 can be specifically for:
According to predefined division limits, along continuous straight runs is divided described image to be identified, obtains a plurality of image blocks.
Optionally, during target behaviour face to be identified in image to be identified, the classification results in taxon 43 can comprise: normal picture piece and abnormal image piece; And the decision rule in classification determining unit 44 can be: while comprising abnormal image piece in the classification results that taxon 43 obtains, image to be identified is defined as to abnormal image, otherwise image to be identified is defined as to normal picture.
Optionally, when image to be identified need to be determined by original image, this device can also comprise:
Receiving element, for receiving original image;
Area definition unit, for by detecting the edge pixel point of the target to be identified of the original image that receiving element receives, determines the area of target to be identified;
The second judging unit, for judging whether the area ratio of target to be identified and described original image is greater than default proportion threshold value;
Image the first determining unit to be identified, for when judgment result is that of the second judging unit is, is defined as image to be identified by original image;
Image the second determining unit to be identified, when the determination result is NO at the second judging unit, determines rule according to default rectangular area, determines the rectangular area that comprises target to be identified, and rectangular area is defined as to image to be identified from original image; Wherein, rectangular area determines that rule meets: make described target to be identified and the area ratio of the rectangular area of determining be greater than described proportion threshold value.
Optionally, the target behaviour face to be identified in image to be identified, and while being determined by original image, this device can also comprise:
Detecting unit, for receiving original image, and detects in original image, whether to comprise people's face;
Boundary rectangle region determining unit while comprising people's face for detect original image when detecting unit, is determined the boundary rectangle region of people's face from original image;
Cutting unit, the region of determining for be partitioned into boundary rectangle region determining unit from original image.
In the situation that above-mentioned image to be identified is determined by original image, division unit 41 can be specifically for:
According to predefined division rule, the region that cutting unit is partitioned into is divided.
The pattern recognition device that the embodiment of the present invention provides, by each image block obtaining after whole image divided, identify respectively and classify, when image block under the unusual part in abnormal image is identified, can not be subject to the impact of the affiliated image block of other normal parts, thereby the accuracy rate of the classification results of each image block can improve greatly, and the accuracy rate of the recognition result of whole the image finally obtaining also can improve, also avoided in prior art due to the feature of the image of the unusual part feature outstanding problem that causes flase drop with respect to whole image simultaneously.
Those skilled in the art should understand, embodiments of the invention can be provided as method, system or computer program.Therefore, the present invention can adopt complete hardware implementation example, implement software example or in conjunction with the form of the embodiment of software and hardware aspect completely.And the present invention can adopt the form that wherein includes the upper computer program of implementing of computer-usable storage medium (including but not limited to magnetic disk memory, CD-ROM, optical memory etc.) of computer usable program code one or more.
The present invention is with reference to describing according to process flow diagram and/or the block scheme of the method for the embodiment of the present invention, equipment (system) and computer program.Should understand can be in computer program instructions realization flow figure and/or block scheme each flow process and/or the flow process in square frame and process flow diagram and/or block scheme and/or the combination of square frame.Can provide these computer program instructions to the processor of multi-purpose computer, special purpose computer, Embedded Processor or other programmable data processing device to produce a machine, the instruction of carrying out by the processor of computing machine or other programmable data processing device is produced for realizing the device in the function of flow process of process flow diagram or a plurality of flow process and/or square frame of block scheme or a plurality of square frame appointments.
These computer program instructions also can be stored in energy vectoring computer or the computer-readable memory of other programmable data processing device with ad hoc fashion work, the instruction that makes to be stored in this computer-readable memory produces the manufacture that comprises command device, and this command device is realized the function of appointment in flow process of process flow diagram or a plurality of flow process and/or square frame of block scheme or a plurality of square frame.
These computer program instructions also can be loaded in computing machine or other programmable data processing device, make to carry out sequence of operations step to produce computer implemented processing on computing machine or other programmable devices, thereby the instruction of carrying out is provided for realizing the step of the function of appointment in flow process of process flow diagram or a plurality of flow process and/or square frame of block scheme or a plurality of square frame on computing machine or other programmable devices.
Although described the preferred embodiments of the present invention, once those skilled in the art obtain the basic creative concept of cicada, can make other change and modification to these embodiment.So claims are intended to all changes and the modification that are interpreted as comprising preferred embodiment and fall into the scope of the invention.
Obviously, those skilled in the art can carry out various changes and modification and not depart from the spirit and scope of the present invention the present invention.Like this, if within of the present invention these are revised and modification belongs to the scope of the claims in the present invention and equivalent technologies thereof, the present invention is also intended to comprise these changes and modification interior.
Claims (10)
1. an image-recognizing method, is characterized in that, comprising:
According to predefined division rule, treat recognition image and divide, obtain a plurality of image blocks; Wherein, described division rule meets: in each image block that makes to obtain, all comprise the pixel that can characterize the clarification of objective to be identified in described image to be identified; And
Described a plurality of image blocks are carried out respectively to feature extraction, obtain the characteristic of each image block;
According to the characteristic of each image block obtaining, and according to described division rule, train the sorter for differentiate between images piece classification obtaining in advance, respectively each image block is classified, obtain the classification results of each image block;
According to the classification results of each image block and default decision rule, determine the classification of described image.
2. the method for claim 1, is characterized in that, the target behaviour face to be identified in described image to be identified; ?
According to predefined division rule, the image to be identified obtaining is divided, obtain a plurality of image blocks and specifically comprise:
According to predefined division limits, along continuous straight runs is divided described image to be identified, obtains a plurality of image blocks.
3. the method for claim 1, is characterized in that, the target behaviour face to be identified in described image to be identified; ?
Described classification results comprises: normal picture piece and abnormal image piece; And
Described decision rule is: while comprising abnormal image piece in described a plurality of image blocks, image to be identified is defined as to abnormal image, otherwise image to be identified is defined as to normal picture.
4. the method as described in as arbitrary in claim 1～3, is characterized in that the target behaviour face to be identified in described image to be identified; ?
Described method also comprises:
Receive original image, and detect in described original image, whether to comprise people's face;
While comprising people's face in detecting described original image, from described original image, determine the boundary rectangle region of described people's face;
From described original image, be partitioned into described region; ?
According to predefined division rule, treat recognition image and divide, specifically comprise:
According to predefined division rule, the described region being partitioned into is divided.
5. the method as described in as arbitrary in claim 1～3, is characterized in that, also comprises:
Receive original image; And
By detecting the edge pixel point of the target to be identified in described original image, determine the area of described target to be identified;
Whether the area ratio that judges described target to be identified and described original image is greater than default proportion threshold value;
Judgment result is that while being, original image is defined as to image to be identified;
When the determination result is NO, according to default rectangular area, determine rule, from described original image, determine the rectangular area that comprises described target to be identified, and described rectangular area is defined as to image to be identified; Wherein, rectangular area determines that rule meets: make described target to be identified and the area ratio of the rectangular area of determining be greater than described proportion threshold value; ?
According to predefined division rule, treat recognition image and divide, specifically comprise:
According to predefined division rule, the image to be identified of determining is divided.
6. a pattern recognition device, is characterized in that, comprising:
Division unit, for according to predefined division rule, treats recognition image and divides, and obtains a plurality of image blocks; Wherein, described division rule meets: in each image block that makes to obtain, all comprise the pixel that can characterize the clarification of objective to be identified in described image to be identified;
Feature extraction unit, carries out respectively feature extraction for described a plurality of image blocks that division unit is obtained, and obtains the characteristic of each image block;
Taxon, characteristic for each image block of obtaining according to feature extraction unit, and according to described division rule, train the sorter for differentiate between images piece classification obtaining in advance, and respectively each image block is classified, obtain the classification results of each image block;
Classification determining unit, for the classification results of each image block of obtaining according to taxon and default decision rule, determines the classification of described image.
7. device as claimed in claim 6, is characterized in that, the target behaviour face to be identified in described image to be identified; ?
Described division unit specifically for:
According to predefined division limits, along continuous straight runs is divided described image to be identified, obtains a plurality of image blocks.
8. device as claimed in claim 6, is characterized in that, the target behaviour face to be identified in described image to be identified; ?
Classification results in taxon comprises: normal picture piece and abnormal image piece; And
Decision rule in classification determining unit is: while comprising abnormal image piece in the classification results that taxon obtains, image to be identified is defined as to abnormal image, otherwise image to be identified is defined as to normal picture.
9. the device as described in as arbitrary in claim 6～8, is characterized in that the target behaviour face to be identified in described image to be identified; ?
Described device also comprises:
Detecting unit, for receiving original image, and detects in described original image, whether to comprise people's face;
Boundary rectangle region determining unit while comprising people's face for detect described original image when detecting unit, is determined the boundary rectangle region of described people's face from described original image;
Cutting unit, for being partitioned into the described region that boundary rectangle region determining unit is determined from described original image; ?
Division unit, specifically for:
According to predefined division rule, the described region that cutting unit is partitioned into is divided.
10. the device as described in as arbitrary in claim 6～8, is characterized in that, described device also comprises:
Receiving element, for receiving original image;
Area definition unit, for by detecting the edge pixel point of the target to be identified of the described original image that receiving element receives, determines the area of described target to be identified;
The second judging unit, for judging whether the area ratio of described target to be identified and described original image is greater than default proportion threshold value;
Image the first determining unit to be identified, for when judgment result is that of the second judging unit is, is defined as image to be identified by original image;
Image the second determining unit to be identified, for at the second judging unit when the determination result is NO, according to default rectangular area, determine rule, from described original image, determine the rectangular area that comprises described target to be identified, and described rectangular area is defined as to image to be identified; Wherein, rectangular area determines that rule meets: make described target to be identified and the area ratio of the rectangular area of determining be greater than described proportion threshold value; ?
Described division unit specifically for:
According to predefined division rule, the image to be identified of determining is divided.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310109749.5A CN104077594B (en) | 2013-03-29 | 2013-03-29 | A kind of image-recognizing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310109749.5A CN104077594B (en) | 2013-03-29 | 2013-03-29 | A kind of image-recognizing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104077594A true CN104077594A (en) | 2014-10-01 |
| CN104077594B CN104077594B (en) | 2018-01-12 |
Family
ID=51598840
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310109749.5A Active CN104077594B (en) | 2013-03-29 | 2013-03-29 | A kind of image-recognizing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104077594B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106203450A (en) * | 2016-07-11 | 2016-12-07 | 国家新闻出版广电总局广播科学研究院 | Based on degree of depth learning framework, image is carried out the object detection method of feature extraction |
| CN107679024A (en) * | 2017-09-11 | 2018-02-09 | 畅捷通信息技术股份有限公司 | The method of identification form, system, computer equipment, readable storage medium storing program for executing |
| CN107729891A (en) * | 2017-12-01 | 2018-02-23 | 旗瀚科技有限公司 | Face characteristic region partitioning method in the case of a kind of non-alignment |
| CN108875336A (en) * | 2017-11-24 | 2018-11-23 | 北京旷视科技有限公司 | The method of face authentication and typing face, authenticating device and system |
| CN109192158A (en) * | 2018-09-21 | 2019-01-11 | 重庆惠科金渝光电科技有限公司 | Control method, display panel and the storage medium of display panel |
| CN109241943A (en) * | 2018-10-09 | 2019-01-18 | 深圳市三宝创新智能有限公司 | Non-alignment face feature extraction method, device, computer equipment and storage medium |
| CN109934305A (en) * | 2019-04-01 | 2019-06-25 | 成都大学 | Image-recognizing method and device based on image recognition model |
| CN110222631A (en) * | 2019-06-04 | 2019-09-10 | 电子科技大学 | Based on deblocking it is parallel cut space arrangement SAR image target recognition method |
| WO2020133072A1 (en) * | 2018-12-27 | 2020-07-02 | Zhejiang Dahua Technology Co., Ltd. | Systems and methods for target region evaluation and feature point evaluation |
| CN112150436A (en) * | 2020-09-23 | 2020-12-29 | 创新奇智(合肥)科技有限公司 | Lipstick inner wall gluing detection method and device, electronic equipment and storage medium |
| CN112561893A (en) * | 2020-12-22 | 2021-03-26 | 平安银行股份有限公司 | Picture matching method and device, electronic equipment and storage medium |
| CN113111925A (en) * | 2021-03-29 | 2021-07-13 | 宁夏新大众机械有限公司 | Feed qualification classification method based on deep learning |
| CN114067370A (en) * | 2022-01-17 | 2022-02-18 | 北京新氧科技有限公司 | Neck shielding detection method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101093539A (en) * | 2007-07-27 | 2007-12-26 | 哈尔滨工程大学 | Matching identification method by extracting characters of vein from finger |
| CN101401730A (en) * | 2008-11-14 | 2009-04-08 | 南京大学 | Fast detecting method for mammary gland tumour shadiness area based on layered structure |
| CN101916442A (en) * | 2010-08-05 | 2010-12-15 | 大连理工大学 | Method for robustly positioning tampered region by utilizing GLCM characteristic |
-
2013
- 2013-03-29 CN CN201310109749.5A patent/CN104077594B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101093539A (en) * | 2007-07-27 | 2007-12-26 | 哈尔滨工程大学 | Matching identification method by extracting characters of vein from finger |
| CN101401730A (en) * | 2008-11-14 | 2009-04-08 | 南京大学 | Fast detecting method for mammary gland tumour shadiness area based on layered structure |
| CN101916442A (en) * | 2010-08-05 | 2010-12-15 | 大连理工大学 | Method for robustly positioning tampered region by utilizing GLCM characteristic |
Non-Patent Citations (1)
| Title |
|---|
| 裴佳佳: "基于子空间的人脸识别技术研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106203450A (en) * | 2016-07-11 | 2016-12-07 | 国家新闻出版广电总局广播科学研究院 | Based on degree of depth learning framework, image is carried out the object detection method of feature extraction |
| CN107679024A (en) * | 2017-09-11 | 2018-02-09 | 畅捷通信息技术股份有限公司 | The method of identification form, system, computer equipment, readable storage medium storing program for executing |
| CN108875336A (en) * | 2017-11-24 | 2018-11-23 | 北京旷视科技有限公司 | The method of face authentication and typing face, authenticating device and system |
| CN107729891A (en) * | 2017-12-01 | 2018-02-23 | 旗瀚科技有限公司 | Face characteristic region partitioning method in the case of a kind of non-alignment |
| CN109192158A (en) * | 2018-09-21 | 2019-01-11 | 重庆惠科金渝光电科技有限公司 | Control method, display panel and the storage medium of display panel |
| CN109241943A (en) * | 2018-10-09 | 2019-01-18 | 深圳市三宝创新智能有限公司 | Non-alignment face feature extraction method, device, computer equipment and storage medium |
| WO2020133072A1 (en) * | 2018-12-27 | 2020-07-02 | Zhejiang Dahua Technology Co., Ltd. | Systems and methods for target region evaluation and feature point evaluation |
| CN109934305A (en) * | 2019-04-01 | 2019-06-25 | 成都大学 | Image-recognizing method and device based on image recognition model |
| CN110222631A (en) * | 2019-06-04 | 2019-09-10 | 电子科技大学 | Based on deblocking it is parallel cut space arrangement SAR image target recognition method |
| CN110222631B (en) * | 2019-06-04 | 2022-03-15 | 电子科技大学 | Data block parallel based cut space arrangement SAR image target identification method |
| CN112150436A (en) * | 2020-09-23 | 2020-12-29 | 创新奇智(合肥)科技有限公司 | Lipstick inner wall gluing detection method and device, electronic equipment and storage medium |
| CN112561893A (en) * | 2020-12-22 | 2021-03-26 | 平安银行股份有限公司 | Picture matching method and device, electronic equipment and storage medium |
| CN113111925A (en) * | 2021-03-29 | 2021-07-13 | 宁夏新大众机械有限公司 | Feed qualification classification method based on deep learning |
| CN114067370A (en) * | 2022-01-17 | 2022-02-18 | 北京新氧科技有限公司 | Neck shielding detection method and device, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104077594B (en) | 2018-01-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104077594A (en) | Image recognition method and device | |
| Wei et al. | Multi-vehicle detection algorithm through combining Harr and HOG features | |
| Vennelakanti et al. | Traffic sign detection and recognition using a CNN ensemble | |
| US9147105B2 (en) | Image recognizing apparatus, image recognizing method, and program | |
| CN101350063B (en) | Method and apparatus for locating human face characteristic point | |
| CN101859382B (en) | License plate detection and identification method based on maximum stable extremal region | |
| CN105608441B (en) | Vehicle type recognition method and system | |
| CN111640089B (en) | Defect detection method and device based on feature map center point | |
| CN103902977A (en) | Face identification method and device based on Gabor binary mode | |
| CN104036284A (en) | Adaboost algorithm based multi-scale pedestrian detection method | |
| CN103455820A (en) | Method and system for detecting and tracking vehicle based on machine vision technology | |
| CN103198493A (en) | Target tracking method based on multi-feature self-adaption fusion and on-line study | |
| CN107093175B (en) | A kind of power battery welding quality detection method | |
| CN109389105B (en) | Multitask-based iris detection and visual angle classification method | |
| CN105069447A (en) | Facial expression identification method | |
| Hechri et al. | Robust road lanes and traffic signs recognition for driver assistance system | |
| KR102105954B1 (en) | System and method for accident risk detection | |
| CN103440035A (en) | Gesture recognition system in three-dimensional space and recognition method thereof | |
| CN102855461A (en) | Method and equipment for detecting fingers in images | |
| Indrabayu et al. | Blob modification in counting vehicles using gaussian mixture models under heavy traffic | |
| Shirazi et al. | Contextual combination of appearance and motion for intersection videos with vehicles and pedestrians | |
| Chen et al. | Automatic detection of traffic lights using support vector machine | |
| Biswas et al. | Detection and classification of speed limit traffic signs | |
| Ganapathy et al. | A Malaysian vehicle license plate localization and recognition system | |
| KR101151435B1 (en) | Apparatus and method of recognizing a face |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |