CN104077594B - A kind of image-recognizing method and device - Google Patents
A kind of image-recognizing method and device Download PDFInfo
- Publication number
- CN104077594B CN104077594B CN201310109749.5A CN201310109749A CN104077594B CN 104077594 B CN104077594 B CN 104077594B CN 201310109749 A CN201310109749 A CN 201310109749A CN 104077594 B CN104077594 B CN 104077594B
- Authority
- CN
- China
- Prior art keywords
- images
- image
- recognized
- block
- identified
- 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.)
- Active
Links
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a kind of image-recognizing method and device, to solve the problems, such as that existing false drop rate is high when image is identified in the prior art.Including：According to division rule set in advance, images to be recognized is divided, obtains multiple images block；Wherein, division rule meets：The pixel for the clarification of objective to be identified that can be characterized in images to be recognized is included in each image block for making to obtain；And feature extraction is carried out respectively to multiple images block, obtain the characteristic of each image block；According to the characteristic of obtained each image block, and the grader for being used to distinguish image block classification for training previously according to division rule to obtain, each image block is classified respectively, obtains the classification results of each image block；According to the classification results of each image block and default decision rule, the classification of described image is determined.
Description
Technical field
The present invention relates to area of pattern recognition, more particularly to a kind of image-recognizing method and device.
Background technology
At present, mode identification technology has obtained more and more extensive application, and it is in safety, finance, man-machine interaction, letter
The numerous areas such as breath, education have a wide range of applications.Existing mode identification technology typically will when carrying out image recognition
Choose suitable feature in itself according to image, preferable recognition effect could be obtained.For example, in face recognition process, typically
It is to input whole image to be identified in face identification device, and image and the image pattern of face identification device to input
Sample image in storehouse is compared, if the characteristic of the whole image and the characteristic of sample image in image pattern storehouse
Mismatch, then the image of the input is defined as abnormal image, and the face in the image of the input is defined as abnormal face.
Wherein, abnormal face generally refers to wear the face that can shelter from face organ's accessories, and normal face is to be directed to abnormal people
For face, all non-abnormal faces are accordingly to be regarded as normal face.
Following defect be present in above-mentioned existing method：False drop rate is higher.Such as the image of the face in identification wear dark glasses
When, as the latter half of the face in image with normal face is, this results in the feature of whole image and normal face
Image feature relatively, and the feature of the image of wear dark glasses part will be than less prominent, so as to will be easy to lead
Flase drop is caused, this abnormal image is judged as normal picture.
The content of the invention
The embodiment of the present invention provides a kind of image-recognizing method and device, to solve to carry out to image in the prior art
The problem of existing false drop rate is higher during identification.
The embodiment of the present invention uses following technical scheme：
A kind of image-recognizing method, including：
According to division rule set in advance, images to be recognized is divided, obtains multiple images block；Wherein, it is described
Division rule meets：The spy for the target to be identified that can be characterized in the images to be recognized is included in each image block for making to obtain
The pixel of sign；And
Feature extraction is carried out respectively to the multiple image block, obtains the characteristic of each image block；
According to the characteristic of obtained each image block, and previously according to the division rule and that trains to obtain is used for
The grader of image block classification is distinguished, each image block is classified respectively, obtains the classification results of each image block；
According to the classification results of each image block and default decision rule, the classification of described image is determined.
A kind of pattern recognition device, including：
Division unit, for according to division rule set in advance, being divided to images to be recognized, obtaining multiple images
Block；Wherein, the division rule meets：Include to characterize in each image block for making to obtain and treated in the images to be recognized
Identify the pixel of clarification of objective；
Feature extraction unit, the multiple image block for being obtained to division unit carry out feature extraction, obtained respectively
The characteristic of each image block；
Taxon, for the characteristic of each image block obtained according to feature extraction unit, and previously according to institute
The grader for being used to distinguish image block classification stated division rule and train to obtain, classifies to each image block, obtains respectively
The classification results of each image block；
Classification determination unit, for the classification results of each image block that are obtained according to taxon and default judge rule
Then, the classification of described image is determined.
The embodiment of the present invention has the beneficial effect that：
The embodiment of the present invention is identified and classified respectively by each image block obtained after being divided to whole image so that
, will not be by the image belonging to other normal segments when the image block belonging to the unusual part in abnormal image is identified
The influence of block, it is relatively more accurate so as to the classification results of each image block, so that the identification knot of the whole image finally given
The accuracy rate of fruit can also improve, avoid in the prior art due to unusual part image feature relative to whole image spy
The problem of sign does not protrude and causes flase drop.
Brief description of the drawings
Fig. 1 is a kind of broad flow diagram for image-recognizing method that the embodiment of the present invention one provides；
Fig. 2 is a kind of particular flow sheet for face identification method that the embodiment of the present invention two provides；
Fig. 3 is the division schematic diagram for the images to be recognized that the embodiment of the present invention two provides；
Fig. 4 is a kind of structural representation for pattern recognition device that the embodiment of the present invention three provides.
Embodiment
In order to solve the problems, such as that existing false drop rate is higher when image is identified in the prior art, the present invention is implemented
Example provides a kind of image recognition scheme.The program is identified respectively by each image block obtained after being divided to whole image
And classify so that, will not be by other normal portions when the image block belonging to the unusual part in abnormal image is identified
The influence of image block belonging to point, so that the classification results of each image block is relatively more accurate, so that what is finally given is whole
The accuracy rate of the recognition result of image can also improve, at the same it also avoid in the prior art due to unusual part image spy
The problem of sign does not protrude relative to the feature of whole image and causes flase drop.
With reference to each accompanying drawing to the main realization principle of technical scheme of the embodiment of the present invention, embodiment and its
The beneficial effect that should be able to reach is explained in detail.
Embodiment one：
As shown in figure 1, for a kind of broad flow diagram of image-recognizing method provided in an embodiment of the present invention, this method includes
Following steps：
Step 11, according to division rule set in advance, images to be recognized is divided, obtains multiple images block；
Specifically, under different application scenarios and image type, there can be different division methods to images to be recognized
Divided.For example when facial image is identified, two image blocks can be divided into the horizontal direction, respectively
For upper half-face image block and lower half-face image block；Or facial image can also be divided into three or four in the horizontal direction
Image block.When other images are identified, vertically images to be recognized can also be divided, even along 45
Angle is oblique that images to be recognized is divided for degree, and this is not restricted for the specific dividing mode embodiment of the present invention, can be according to
Actual demand sets itself.
Wherein, division rule should meet：Being included in each image block for making to obtain to characterize in images to be recognized
The pixel of clarification of objective to be identified.
In addition, images to be recognized is directly obtained or determined by original image.Wherein, when connecing
When the image received is original image, a kind of mode for determining images to be recognized can specifically include following step：
First, by detecting the edge pixel point of the target to be identified in the original image received, mesh to be identified is determined
Target area；
Then, judge whether the area ratio of target to be identified and original image is more than default proportion threshold value；
When judged result is to be, original image is defined as images to be recognized；
When judged result is no, rule is determined according to default rectangular area, determines to include from original image and waits to know
The rectangular area of other target, and rectangular area is defined as images to be recognized, wherein, rectangular area determines that rule should meet：
So that target to be identified and the area ratio for the rectangular area determined are more than above-mentioned proportion threshold value.
When target to be identified is face, images to be recognized can be determined according to above-mentioned determination mode, can also be according to
The mode of being identified below is determined, and is specifically included：
First, original image is received, and detects in the original image whether include face；
Then, when detecting to include face in original image, where the extraneous rectangle that face is determined from original image
Region；
Finally, above-mentioned zone is partitioned into from original image, and is defined as images to be recognized.
It should be noted that mentioned above determine the rectangular area determined of rule according to rectangular area, and from original
The boundary rectangle region for the face being partitioned into beginning image, can first scale it into the image of default size, then will enter
The image obtained after row scaling is defined as images to be recognized.
Step 12, feature extraction is carried out respectively to obtained multiple images block, obtains the characteristic of each image block；
Wherein it is possible to according to edge orientation histogram, local binary（Local Binary Pattern, LBP), chi
Spend invariant features conversion（Scale Invariant Feature Transform, SIFT）With acceleration robust features (Speeded-
Up Robust Feature, SURF) etc. feature obtain the characteristic of each image block.
Step 13, according to the characteristic of obtained each image block, and previously according to division rule train what is obtained
For distinguishing the grader of image block classification, each image block is classified respectively, obtains the classification results of each image block.
Step 14, according to the classification results of each image block and default decision rule, the classification of the image is determined.
The embodiment of the present invention is identified and classified respectively by each image block obtained after being divided to whole image so that
, will not be by the image belonging to other normal segments when the image block belonging to the unusual part in abnormal image is identified
The influence of block, it is relatively more accurate so as to the classification results of each image block, so that the identification knot of the whole image finally given
The accuracy rate of fruit can also improve, while it also avoid in the prior art because the feature of the image of unusual part is schemed relative to whole
The problem of feature of picture does not protrude and causes flase drop.
Embodiment two：
The embodiment of the present invention two image-recognizing method in embodiment one will be described in detail with reference to practical application, this hair
Illustrated in bright embodiment exemplified by above-mentioned image-recognizing method is applied into recognition of face, specific application scenario can be
The occasions such as ATM monitoring video automatic alarms.When someone is withdrawing cash, face can be classified automatically, be according to it then
No exception chooses whether to alarm, and can be linked with ATM, and when identifying that face is abnormal face, ATM is without telling paper money.
As shown in Fig. 2 it is a kind of particular flow sheet of face identification method provided in an embodiment of the present invention.
Step 21, original image is received；
Step 22, detect in original image whether include face；When detecting to include face in original image, step is performed
Rapid 23, the original image is directly otherwise defined as abnormal image.
Step 23, the boundary rectangle of face is determined from original image, and by the boundary rectangle region from original
Split in image as images to be recognized；
Wherein, the determination of above-mentioned images to be recognized can also enter according to the another way introduced in the step 11 of embodiment one
Row determines.
Step 24, according to division rule set in advance, pair images to be recognized determined divides in the horizontal direction, obtains
To upper half-face image block and lower half-face image block；
Wherein, division rule should meet：Make the upper half-face image block that is obtained after division and wrapped in lower half-face image block
The pixel of feature containing the face that can be characterized in images to be recognized.
For example the height and width of the images to be recognized determined are respectively H and W（I.e. high H pixel, wide W pixel
Point）, then can be divided, the images to be recognized is placed in reference axis, as shown in figure 3, x-axis side with division limits YO=H/2
To width is represented, y-axis direction represents height, then by x-axis from 0 to W-1, part of the y-axis from 0 to YO-1 is as lower half-face image
Block, by x-axis from 0 to W-1, part of the y-axis from YO to H-1 is as upper half-face image block.
Step 25, feature extraction is carried out respectively to upper half-face image block and lower half-face image block, obtains half-face image block
Characteristic and lower half-face image block characteristic；
Wherein, for upper half-face image block, by taking the feature for extracting edge orientation histogram as an example, detailed process is as follows：
Rim detection is carried out to obtained upper half-face image block, obtains the edge intensity value computing and edge direction of each pixel
Value.
For example this edge detection operator of Sobel operators can be used to carry out convolution operation to upper half-face image block, so as to
Obtain the Sobel marginal values in x-axis and y-axis direction.Wherein, can be successively for the Sobel operators of x-axis and y-axis：
The edge intensity value computing of each pixel on this in half-face image block is the absolute value and y of the Sobel marginal values of x-axis
The absolute value sum of the Sobel marginal values of axle.
The computational methods of the edge direction values of each pixel on this in half-face image block are as follows：
Wherein, formula（1）In, Dir is the edge direction values of pixel, and YSobel is the Sobel marginal values of y-axis,
XSobel is the Sobel marginal values of x-axis.
Wherein, the size of the edge intensity value computing of pixel determines whether the pixel is edge pixel point, if its edge is strong
Angle value is too small, then this pixel is not just edge pixel point, and can does not count when carrying out edge orientation histogram statistics.Cause
This, is more than default pixel edge intensity value computing in the embodiment of the present invention to edge intensity value computing（Hereinafter referred to as predetermined threshold value）Picture
Vegetarian refreshments carries out the statistics of edge orientation histogram, and the edge intensity value computing that will be greater than each pixel of predetermined threshold value is mapped as edge
Numerical value in direction histogram, the calculation formula of use are as follows：
Wherein, formula（2）In, Grad is the edge intensity value computing of pixel in image block, and SobelThres is default picture
Vegetarian refreshments edge intensity value computing, Hist (Dir) are the numerical value in edge orientation histogram.
Finally, the edge orientation histogram of pixel, determines its characteristic in the upper half-face image block obtained according to counting
According to.
Likewise, the feature extraction for lower half-face image block is consistent with said process, therefore not to repeat here.
It should be noted that the method for feature extraction has a lot, it is contemplated that the embodiment of the present invention is that face is identified.
And under normal circumstances, when face is identified, when face has tilted or turns to a little（It is exactly face little on one side）
When, just easily lead to flase drop.And the method can of this feature of extraction edge orientation histogram is used to avoid above-mentioned face
The problem that during inclination.Because the edge orientation histogram extracted when face has tilted and the side of normal face
Edge direction histogram is consistent.Therefore, the preferred method for extracting edge orientation histogram in the embodiment of the present invention, also may be used certainly
Other feature extracting methods are used to be not limited to the above method, specifically can voluntarily be selected according to actual conditions.
Step 26, according to the characteristic of obtained each image block, and previously according to division rule train what is obtained
For distinguishing the grader of normal picture block and abnormal image block, each image block is classified respectively, obtains each image block
Classification results.
Wherein, training the specific method of grader can be：Choose certain amount（Two or three hundred）Image pattern,
Then the image pattern of selection is divided and feature extraction respectively, respectively obtains the characteristic of some upper half-face image blocks
According to the characteristic with lower half-face image block, as two classifications, it is respectively trained out for half-face image block on one
Grader and a grader for being directed to lower half-face image block.Have much on the specific training method of grader, can basis
Actual conditions are selected, and will not be described in detail herein in the embodiment of the present invention.
Step 27, according to upper half-face image block and the classification results of lower half-face image block, whether judge wherein comprising abnormal
Image block, if comprising, perform step 28, otherwise perform step 29.
Step 28, images to be recognized is defined as abnormal image.
Step 29, images to be recognized is defined as normal picture.
It is identified and classifies respectively by each image block obtained after being divided to whole image in the embodiment of the present invention, makes
Obtain when the image block belonging to the unusual part in abnormal image is identified, will not be by the figure belonging to other normal segments
As the influence of block, so that the accuracy rate of the classification results of each image block can greatly improve, and the knowledge of the whole image finally given
The accuracy rate of other result can also improve, at the same it also avoid in the prior art due to unusual part image feature relative to whole
The feature of image is not prominent and the problem of cause flase drop.
Embodiment three：
The embodiment of the present invention three additionally provides a kind of pattern recognition device, and the structural representation of the device is as shown in figure 4, bag
Include：
Division unit 41, for according to division rule set in advance, being divided to images to be recognized, obtaining multiple figures
As block；Wherein, division rule meets：Comprising the mesh to be identified that can be characterized in images to be recognized in each image block for making to obtain
The pixel of target feature；
Feature extraction unit 42, the multiple images block for being obtained to division unit 41 carry out feature extraction, obtained respectively
The characteristic of each image block；
Taxon 43, for the characteristic of each image block obtained according to feature extraction unit 42, and advance root
That trains to obtain according to division rule is used to distinguish the grader of image block classification, and each image block is classified respectively, obtained
The classification results of each image block；
Classification determination unit 44, for the classification results of each image block that are obtained according to taxon 43 and default sentence
Set pattern then, determines the classification of the image.
Wherein, when the target to be identified in images to be recognized is face, division unit 41 can be specifically used for：
According to division limits set in advance, the images to be recognized is divided in the horizontal direction, obtains multiple figures
As block.
Optionally, when the target to be identified in images to be recognized is face, classification results in taxon 43 can be with
Including：Normal picture block and abnormal image block；And the decision rule in classification determination unit 44 can be：When taxon 43
When abnormal image block is included in obtained classification results, images to be recognized is defined as abnormal image, otherwise by images to be recognized
It is defined as normal picture.
Optionally, when images to be recognized needs to be determined by original image, the device can also include：
Receiving unit, for receiving original image；
Area determining unit, the edge picture for the target to be identified in the original image by detecting receiving unit reception
Vegetarian refreshments, determine the area of target to be identified；
Second judging unit, it is default whether the area ratio for judging target to be identified and the original image is more than
Proportion threshold value；
The determining unit of images to be recognized first, for the second judging unit judged result for be when, by original image
It is defined as images to be recognized；
The determining unit of images to be recognized second, for the second judging unit judged result for it is no when, according to default
Rectangular area determines rule, the rectangular area for including target to be identified is determined from original image, and rectangular area is defined as
Images to be recognized；Wherein, rectangular area determines that rule meets：So that the face of the target to be identified and the rectangular area determined
Product ratio is more than the proportion threshold value.
Optionally, the target to be identified in images to be recognized is face, and when being determined by original image, the device may be used also
With including：
Detection unit, for receiving original image, and detect in original image whether include face；
Boundary rectangle region determining unit, for when detection unit detects to include face in original image, from
The boundary rectangle region of face is determined in original image；
Cutting unit, the region determined for being partitioned into boundary rectangle region determining unit from original image.
In the case where above-mentioned images to be recognized is determined by original image, division unit 41 can be specifically used for：
According to division rule set in advance, the region that cutting unit is partitioned into is divided.
Pattern recognition device provided in an embodiment of the present invention, distinguished by each image block obtained after being divided to whole image
It is identified and classifies so that, will not be by it when the image block belonging to the unusual part in abnormal image is identified
The influence of image block belonging to his normal segments, so as to which the accuracy rate of the classification results of each image block can greatly improve, and it is final
The accuracy rate of the recognition result of whole obtained image can also improve, while it also avoid in the prior art due to unusual part
The problem of feature of image does not protrude relative to the feature of whole image and causes flase drop.
It should be understood by those skilled in the art that, embodiments of the invention can be provided as method, system or computer program
Product.Therefore, the present invention can use the reality in terms of complete hardware embodiment, complete software embodiment or combination software and hardware
Apply the form of example.Moreover, the present invention can use the computer for wherein including computer usable program code in one or more
Usable storage medium（Including but not limited to magnetic disk storage, CD-ROM, optical memory etc.）The computer program production of upper implementation
The form of product.
The present invention is with reference to method according to embodiments of the present invention, equipment（System）And the flow of computer program product
Figure and/or block diagram describe.It should be understood that can be by every first-class in computer program instructions implementation process figure and/or block diagram
Journey and/or the flow in square frame and flow chart and/or block diagram and/or the combination of square frame.These computer programs can be provided
The processors of all-purpose computer, special-purpose computer, Embedded Processor or other programmable data processing devices is instructed to produce
A raw machine so that produced by the instruction of computer or the computing device of other programmable data processing devices for real
The device for the function of being specified in present one flow of flow chart or one square frame of multiple flows and/or block diagram or multiple square frames.
These computer program instructions, which may be alternatively stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that the instruction being stored in the computer-readable memory, which produces, to be included referring to
Make the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one square frame of block diagram or
The function of being specified in multiple square frames.
These computer program instructions can be also loaded into computer or other programmable data processing devices so that counted
Series of operation steps is performed on calculation machine or other programmable devices to produce computer implemented processing, so as in computer or
The instruction performed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in individual square frame or multiple square frames.
Although preferred embodiments of the present invention have been described, but those skilled in the art once know basic creation
Property concept, then can make other change and modification to these embodiments.So appended claims be intended to be construed to include it is excellent
Select embodiment and fall into having altered and changing for the scope of the invention.
Obviously, those skilled in the art can carry out the essence of various changes and modification without departing from the present invention to the present invention
God and scope.So, if these modifications and variations of the present invention belong to the scope of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to comprising including these changes and modification.
Claims (8)
- A kind of 1. image-recognizing method, it is characterised in that including：According to division rule set in advance, images to be recognized is divided, obtains multiple images block；Wherein, the division Rule meets：Comprising the clarification of objective to be identified that can be characterized in the images to be recognized in each image block for making to obtain Pixel；AndFeature extraction is carried out respectively to the multiple image block, obtains the characteristic of each image block；According to the characteristic of obtained each image block, and that trains previously according to the division rule to obtain be used to distinguish The grader of image block classification, each image block is classified respectively, obtain the classification results of each image block, wherein, described point Class result includes：Normal picture block and abnormal image block；When including abnormal image block in the multiple image block, the images to be recognized is defined as abnormal image, otherwise will The images to be recognized is defined as normal picture.
- 2. the method as described in claim 1, it is characterised in that the target to be identified in the images to be recognized is face；ThenAccording to division rule set in advance, the images to be recognized of acquisition is divided, multiple images block is obtained and specifically includes：According to division limits set in advance, the images to be recognized is divided in the horizontal direction, obtains multiple images block.
- 3. the method as described in claim 1~2 is any, it is characterised in that the target to be identified in the images to be recognized is Face；ThenMethods described also includes：Original image is received, and detects in the original image whether include face；When detecting to include face in the original image, the boundary rectangle institute of the face is determined from the original image In region；The region is partitioned into from the original image；ThenAccording to division rule set in advance, images to be recognized is divided, specifically included：According to division rule set in advance, the region being partitioned into is divided.
- 4. the method as described in claim 1~2 is any, it is characterised in that also include：Receive original image；AndBy detecting the edge pixel point of the target to be identified in the original image, the area of the target to be identified is determined；Judge whether the target to be identified and the area ratio of the original image are more than default proportion threshold value；When judged result is to be, then original image is defined as images to be recognized；When judged result is no, rule is determined according to default rectangular area, is determined from the original image comprising described in The rectangular area of target to be identified, and the rectangular area is defined as images to be recognized；Wherein, rectangular area determines that rule is full Foot：So that the target to be identified and the area ratio for the rectangular area determined are more than the proportion threshold value；ThenAccording to division rule set in advance, images to be recognized is divided, specifically included：According to division rule set in advance, the images to be recognized determined is divided.
- A kind of 5. pattern recognition device, it is characterised in that including：Division unit, for according to division rule set in advance, being divided to images to be recognized, obtaining multiple images block； Wherein, the division rule meets：Include to characterize in each image block for making to obtain and wait to know in the images to be recognized The pixel of other clarification of objective；Feature extraction unit, the multiple image block for being obtained to division unit carry out feature extraction, obtain each figure respectively As the characteristic of block；Taxon, for the characteristic of each image block obtained according to feature extraction unit, and previously according to described stroke Divider then trains the obtained grader for being used to distinguish image block classification, and each image block is classified respectively, obtains each figure As the classification results of block, wherein, the classification results include：Normal picture block and abnormal image block；Classification determination unit, will be described to be identified for when including abnormal image block in the classification results that taxon obtains Image is defined as abnormal image, and the images to be recognized otherwise is defined as into normal picture.
- 6. device as claimed in claim 5, it is characterised in that the target to be identified in the images to be recognized is face；ThenThe division unit is specifically used for：According to division limits set in advance, the images to be recognized is divided in the horizontal direction, obtains multiple images block.
- 7. the device as described in claim 5~6 is any, it is characterised in that the target to be identified in the images to be recognized is Face；ThenDescribed device also includes：Detection unit, for receiving original image, and detect in the original image whether include face；Boundary rectangle region determining unit, for when detection unit detects to include face in the original image, from The boundary rectangle region of the face is determined in the original image；Cutting unit, the area determined for being partitioned into boundary rectangle region determining unit from the original image Domain；ThenDivision unit, it is specifically used for：According to division rule set in advance, the region being partitioned into cutting unit divides.
- 8. the device as described in claim 5~6 is any, it is characterised in that described device also includes：Receiving unit, for receiving original image；Area determining unit, the edge picture for the target to be identified in the original image by detecting receiving unit reception Vegetarian refreshments, determine the area of the target to be identified；Second judging unit, it is default whether the area ratio for judging the target to be identified and the original image is more than Proportion threshold value；The determining unit of images to be recognized first, for when the judged result of the second judging unit is is, original image to be determined For images to be recognized；The determining unit of images to be recognized second, for the second judging unit judged result for it is no when, according to default rectangle Region determines rule, determines to include the rectangular area of the target to be identified from the original image, and by the rectangle region Domain is defined as images to be recognized；Wherein, rectangular area determines that rule meets：So that the target to be identified and the rectangle determined The area ratio in region is more than the proportion threshold value；ThenThe division unit is specifically used for：According to division rule set in advance, the images to be recognized determined 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 CN104077594A (en) | 2014-10-01 |
| CN104077594B true 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) |
Families Citing this family (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 |
| CN107679024B (en) * | 2017-09-11 | 2023-04-18 | 畅捷通信息技术股份有限公司 | Method, system, computer device and readable storage medium for identifying table |
| 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 of display panel, display panel and storage medium |
| 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 |
| CN111435458A (en) * | 2019-01-11 | 2020-07-21 | 罗伯特·博世有限公司 | Method and apparatus for creating and verifying anti-counterfeiting marks |
| CN109934305A (en) * | 2019-04-01 | 2019-06-25 | 成都大学 | Image recognition method and device based on image recognition model |
| CN110222631B (en) * | 2019-06-04 | 2022-03-15 | 电子科技大学 | Target recognition method of tangential space arrangement SAR image based on data block and parallel |
| 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 |
| CN112561893B (en) * | 2020-12-22 | 2024-09-06 | 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 |
| CN114067370B (en) * | 2022-01-17 | 2022-06-21 | 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 | A rapid detection method for suspicious areas of breast masses based on hierarchical structure |
| CN101916442A (en) * | 2010-08-05 | 2010-12-15 | 大连理工大学 | A robust localization method for tampered images using GLCM features |
-
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 | A rapid detection method for suspicious areas of breast masses based on hierarchical structure |
| CN101916442A (en) * | 2010-08-05 | 2010-12-15 | 大连理工大学 | A robust localization method for tampered images using GLCM features |
Non-Patent Citations (1)
| Title |
|---|
| 基于子空间的人脸识别技术研究;裴佳佳;《中国优秀硕士学位论文全文数据库》;20100215(第2期);第5.3以及5.5节 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104077594A (en) | 2014-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104077594B (en) | A kind of image-recognizing method and device | |
| US9008365B2 (en) | Systems and methods for pedestrian detection in images | |
| CN104008399B (en) | The recognition methodss of the gauge pointer shake based on support vector machine in a kind of instrument detection | |
| CN103034852B (en) | The detection method of particular color pedestrian under Still Camera scene | |
| CN105608441B (en) | Vehicle type recognition method and system | |
| CN107273832B (en) | License plate recognition method and system based on integral channel characteristics and convolutional neural network | |
| CN105740910A (en) | Vehicle object detection method and device | |
| CN103971126A (en) | Method and device for identifying traffic signs | |
| CN102930553A (en) | Method and device for identifying objectionable video content | |
| CN111242899B (en) | Image-based flaw detection method and computer-readable storage medium | |
| CN103455820A (en) | Method and system for detecting and tracking vehicle based on machine vision technology | |
| CN110929635B (en) | Fake face video detection method and system based on facial intersection and comparison under trust mechanism | |
| CN114067186B (en) | Pedestrian detection method and device, electronic equipment and storage medium | |
| CN103632159A (en) | Method and system for training classifier and detecting text area in image | |
| CN106331746B (en) | Method and apparatus for identifying watermark location in video file | |
| CN107240086A (en) | A kind of fabric defects detection method based on integration nomography | |
| CN103413149A (en) | Method for detecting and identifying static target in complicated background | |
| CN105512662A (en) | Detection method and apparatus for unlicensed vehicle | |
| CN106407904A (en) | Bang zone determining method and device | |
| CN113920434A (en) | Image reproduction detection method, device and medium based on target | |
| CN102496030A (en) | Identification method and identification device for dangerous targets in power monitoring system | |
| Boonsim | Racing bib number localization on complex backgrounds | |
| CN106943116B (en) | A kind of infant eyesight automatic testing method | |
| Piérard et al. | A probabilistic pixel-based approach to detect humans in video streams | |
| GB2522259A (en) | A method of object orientation detection |
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 |