CN108549836A - Reproduction detection method, device, equipment and the readable storage medium storing program for executing of photo - Google Patents
Reproduction detection method, device, equipment and the readable storage medium storing program for executing of photo Download PDFInfo
- Publication number
- CN108549836A CN108549836A CN201810196019.6A CN201810196019A CN108549836A CN 108549836 A CN108549836 A CN 108549836A CN 201810196019 A CN201810196019 A CN 201810196019A CN 108549836 A CN108549836 A CN 108549836A
- Authority
- CN
- China
- Prior art keywords
- image
- photo
- key position
- obtains
- reproduction
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Abstract
The present invention relates to a kind of reproduction detection methods of photo, including：The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each key position in photo；Extract the characteristic value in each key position；Characteristic value in each key position is normalized, and merges each normalized characteristic value, obtains fusion feature value；Fusion feature value input grader is subjected to classification and obtains classification results；If the classification results are reproduction class, confirm that the photo is reproduction.The reproduction detection method of the above human face photo, is subdivided into key area by face first, then to the key area carry out feature extraction, to make human face photo reproduction detection method precision higher.The present invention also provides a kind of detection device of reproduction, equipment and readable storage medium storing program for executing.
Description
Technical field
The present invention relates to field of image recognition, more particularly to the reproduction detection method of photo, device, equipment and readable deposit
Storage media.
Background technology
The application of recognition of face at present is more and more extensive, however some criminals utilize the loophole of face recognition technology,
System deception is carried out using face certificate photo reproduction, to face identification system of out-tricking, obtains the illegal permission of system.Therefore, such as
What detection human face photo whether become the hot issue of this field by reproduction.
Traditional reproduction detection method has the reproduction detection method based on video, and it is attached that face in video image is mainly utilized
The temporal information of band, i.e., with the variation of time, certain variation can occur for the facial expression of face, and these variations are inevitable
Lead to the variation of characteristics of image.However, the slight change of facial expression will not occur with the variation of time for reproduction from a photograph, because
This can carry out biological living identification using this category feature.But the reproduction detection method based on video needs user to coordinate,
It is inconvenient to use.
Therefore, there is the reproduction detection method based on single photo, still, detected currently based on the reproduction of single photo
Face is usually carried out feature extraction by method as a whole, due to the different zones of face have the characteristics that it is different, this
The accuracy of kind method detection is relatively low.
Invention content
Based on this, it is necessary to which the relatively low problem of the accuracy that is detected to human face photo reproduction for conventional method provides one
Reproduction detection method, device, equipment and the readable storage medium storing program for executing of kind photo.
A kind of reproduction detection method of photo, wherein the method includes：
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each pass in photo
Key position；
Extract the characteristic value in each key position；
Characteristic value in each key position is normalized, and merges each normalized characteristic value, is obtained
Obtain fusion feature value；
Fusion feature value input grader is subjected to classification and obtains classification results；
If the classification results are reproduction class, confirm that the photo is reproduction.
The reproduction detection method of the above human face photo, is subdivided into key area by face first, then to the key area
Domain carry out feature extraction, to make human face photo reproduction detection method precision higher.
As a kind of embodiment, wherein described that the photo is converted to two-value picture, and the connection based on bianry image
Each key position that domain algorithm obtains in photo includes：
Judge in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, executing, and the connected domain algorithm based on bianry image is obtained and shone
Each key position in piece.
As a kind of embodiment, wherein the photo is converted to two-value picture, and the connected domain based on bianry image is calculated
The key position that method obtains in image to be detected includes：
Color space conversion is carried out to input picture, obtains color image；
By the color image binaryzation, binary image is obtained；
The binary image is vertically mapped, the face area in the binary image is obtained；
The face area is subjected to gradation conversion, obtains the gray level image of the face area；
Image convolution is carried out to the gray level image, and binaryzation is carried out to the image after convolution, obtains binaryzation convolution
Image；
By connective region search algorithm, the key position is extracted from the binaryzation convolved image.
As a kind of embodiment, wherein key position includes nose, eye and oral area；
The step of feature in the extraction key position includes：
Nose feature is extracted using gray level co-occurrence matrixes；
Eye feature is extracted using LBP algorithms；
Oral area feature is extracted using wavelet transformation.
A kind of reproduction detection device of photo, wherein described device includes：
Key position acquisition module, for the photo to be converted to two-value picture, and the connected domain based on bianry image
Algorithm obtains each key position in photo；
Characteristics extraction module, for extracting the feature in each key position；
Characteristic value Fusion Module for the characteristic value in each key position to be normalized, and merges each
Normalized characteristic value obtains fusion feature value；
Classification results acquisition module obtains classification results for fusion feature value input grader to be carried out classification；
Reproduction judgment module confirms that the photo is reproduction if being reproduction class for the classification results.
The reproduction detection device of the above human face photo, is subdivided into key area by face first, then to the key area
Domain carry out feature extraction, to make human face photo reproduction detection method precision higher.
As one embodiment, wherein the key position acquisition module includes：
Face datection unit, for judging in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, being continued to execute by key position acquisition module, and is based on two-value
The connected domain algorithm of image obtains each key position in photo.
As one embodiment, wherein the key position acquisition module includes：
Color image acquiring unit obtains color image for carrying out color space conversion to input picture；
Binary image acquiring unit, for by the color image binaryzation, obtaining binary image；
Face area acquiring unit obtains the binary image for vertically being mapped the binary image
In face area；
Gray level image acquiring unit obtains the ash of the face area for the face area to be carried out gradation conversion
Spend image；
Binaryzation convolved image unit, for the gray level image carry out image convolution, and to the image after convolution into
Row binaryzation obtains binaryzation convolved image；
Key position extraction unit, for by connective region search algorithm, institute to be extracted from the binaryzation convolved image
State key position.
As one embodiment, wherein key position includes nose, eye and oral area；
The characteristics extraction module includes：
Nose feature extraction unit, for extracting nose feature using gray level co-occurrence matrixes；
Eye feature extraction unit, for extracting eye feature using LBP algorithms；
Oral area feature extraction unit, for extracting oral area feature using wavelet transformation.
A kind of computer equipment, including memory, processor and storage can be run on a memory and on a processor
Computer program, wherein when the processor executes described program, realize following steps：
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each pass in photo
Key position；
Extract the feature in each key position；
Characteristic value in each key position is normalized, and merges each normalized characteristic value, is obtained
Obtain fusion feature value；
Fusion feature value input grader is subjected to classification and obtains classification results；
If the classification results are reproduction class, confirm that the photo is reproduction.
The method that computer program in the above computer equipment is realized when being executed by processor, it is first that face is thin
It is divided into key area, the extraction of feature is then carried out to the key area, to makes the reproduction detection method of human face photo
Precision higher.
As one embodiment, wherein the photo is converted to two-value picture by described performed by processor, and is based on
The connected domain algorithm of bianry image obtains each key position in photo, including：
Judge in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, executing, and the connected domain algorithm based on bianry image is obtained and shone
Each key position in piece.
As one embodiment, wherein the photo is converted to two-value picture performed by processor, and is based on two-value
The key position that the connected domain algorithm of image obtains in image to be detected includes：
Color space conversion is carried out to input picture, obtains color image；
By the color image binaryzation, binary image is obtained；
The binary image is vertically mapped, the face area in the binary image is obtained；
The face area is subjected to gradation conversion, obtains the gray level image of the face area；
Image convolution is carried out to the gray level image, and binaryzation is carried out to the image after convolution, obtains binaryzation convolution
Image；
By connective region search algorithm, the key position is extracted from the binaryzation convolved image.
As one embodiment, wherein the key position performed by processor includes nose, eye and oral area；
The step of feature in the extraction key position includes：
Nose feature is extracted using gray level co-occurrence matrixes；
Eye feature is extracted using LBP algorithms；
Oral area feature is extracted using wavelet transformation.
A kind of readable storage medium storing program for executing, the readable storage medium storing program for executing are stored with computer program, and the computer program is located
Reason device realizes following steps when executing：
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each pass in photo
Key position；
Extract the feature in each key position；
Characteristic value in each key position is normalized, and merges each normalized characteristic value, is obtained
Obtain fusion feature value；
Fusion feature value input grader is subjected to classification and obtains classification results；
If the classification results are reproduction class, confirm that the photo is reproduction.
The method that the computer program of the above readable storage medium storing program for executing is realized when being executed by processor, it is first that face is thin
It is divided into key area, the extraction of feature is then carried out to the key area, to makes the reproduction detection method of human face photo
Precision higher.
As one embodiment, wherein can the computer program of storage medium storage realize when being executed by processor
It is described that the photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each key portion in photo
Position, including：
Judge in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, executing, and the connected domain algorithm based on bianry image is obtained and shone
Each key position in piece.
As one embodiment, wherein can the computer program of storage medium storage realize when being executed by processor
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains the crucial portion in image to be detected
Position includes：
Color space conversion is carried out to input picture, obtains color image；
By the color image binaryzation, binary image is obtained；
The binary image is vertically mapped, the face area in the binary image is obtained；
The face area is subjected to gradation conversion, obtains the gray level image of the face area；
Image convolution is carried out to the gray level image, and binaryzation is carried out to the image after convolution, obtains binaryzation convolution
Image；
By connective region search algorithm, the key position is extracted from the binaryzation convolved image.
As one embodiment, wherein can the computer program of storage medium storage realize when being executed by processor
In method, the key position includes nose, eye and oral area；
The step of feature in the extraction key position includes：
Nose feature is extracted using gray level co-occurrence matrixes；
Eye feature is extracted using LBP algorithms；
Oral area feature is extracted using wavelet transformation.
Description of the drawings
Fig. 1 is the flow chart for the reproduction detection method that one of which embodiment provides；
Fig. 2 is the partial process view for the reproduction detection method that one of which embodiment provides；
Fig. 3 is the partial process view for the reproduction detection method that one of which embodiment provides；
Fig. 4 is the schematic diagram of the calculating process for the LBP values that one of embodiment provides；
Fig. 5 is the structural schematic diagram of the reproduction detection device for the photo that one of embodiment provides.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, right with reference to the accompanying drawings and embodiments
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention.
Referring to Fig. 1, Fig. 1 is the flow chart for the reproduction detection method that one of which embodiment provides.
The photo is converted to two-value picture by step S102, and the connected domain algorithm based on bianry image obtains photo
In each key position.
Specifically, obtain key position in image to be detected first, the key position may include nose, eye,
At least one of mouth.
Specifically, face key position detection can be carried out by information such as colors, and obtains key position.First will
Photo is converted to two-value picture, and then the connected domain algorithm based on bianry image obtains the key position in photo again.
Step S104 extracts the feature in each key position；
Specifically, due to each key position the characteristics of, is different, and the extraction of characteristic value is carried out for different parts, can
To improve the precision of human face photo reproduction detection.
Specifically, the methods of geometric method, modelling, signal processing method may be used to put forward the characteristic value of key position
It takes.Gray level co-occurrence matrixes algorithm, LBP algorithms or Wavelet Transformation Algorithm pair can also be respectively adopted according to different key positions
The characteristic value of the key position extracts.For example, gray level co-occurrence matrixes extraction nose feature may be used, may be used
LBP algorithms extract eye feature, and wavelet transformation extraction oral area feature may be used.
Characteristic value in each key position is normalized step S106, and merges each normalized spy
Value indicative obtains fusion feature value.
Specifically, the feature of extraction is merged, can be first returned the characteristic value of the key position of S104 extractions
One changes, and the normalization refers to, using any one position in the key position as standard, the characteristic value at other positions being pressed
It is converted according to this standard.Then splice to obtain fusion feature by the way of being added, the pass that can also first extract S104
Then the feature normalization at key position gives different weights according to specific requirements, is merged by way of weighting summation
Feature.It is appreciated that the fusion feature needs to embody the feature in each key position, but the feature of each different parts
Different weights can be set according to specific requirements, to embody the importance of the feature of different parts in fusion feature not
Together.
Fusion feature value input grader is carried out classification and obtains classification results by step S108；
Specifically, support vector machines (SVM) or other graders can be used to be detected.Some training figures are used first
As being trained, training pattern is generated.Then in conjunction with the training pattern, pass through support vector machines (SVM) or other graders
Classify to the fusion feature of input, obtains classification results, and determine whether the photo is reproduction according to classification results.
Step S110 confirms that the photo is reproduction if the classification results are reproduction class.
Specifically, classification results can be reproduction class and non-reproduction class, if the classification results are reproduction class, institute
It is reproduction from a photograph to state photo.Correspondingly, if classification results are non-reproduction class, it is described to judge that the photo is not reproduction
Photo.The reproduction detection method of the above human face photo, is subdivided into key area by face first, then to the key area into
The extraction of row feature, to make human face photo reproduction detection method precision higher.
The reproduction detection method of the above human face photo, is subdivided into key area by face first, then to the key area
Domain carry out feature extraction, to make human face photo reproduction detection method precision higher.
As the reproduction detection method that one of which embodiment provides, the crucial portion obtained in image to be detected
Position the step of include：Judge in the image to be detected whether to include face information；If so, executing in image to be detected
The step of obtaining key position；If it is not, then terminating this detection.
Specifically, the face in image is detected using Face datection algorithm, directly excludes the face being not present photograph
Piece.The Face datection algorithm can be according to specific requirements using the recognizer of feature based point, based on the knowledge of neural network
Other algorithm etc..
The reproduction detection method of the above human face photo increases before the step of obtaining the key position in image to be detected
The process of one recognition of face directly excludes the photo there is no face, further improves the efficiency of reproduction detection.
As the reproduction detection method that one of which embodiment provides, the key position includes nose, eye, mouth
Portion.
Specifically, in order to obtain higher accuracy of detection, the key position may include nose, eye, oral area this three
A key position.
The reproduction detection method of the above human face photo, while using nose, eye, oral area these three key areas, then
To the key area carry out feature extraction, to make human face photo reproduction detection method precision higher.
It is that one of which is real also referring to Fig. 2, Fig. 2 as the reproduction detection method that one of which embodiment provides
The partial process view for the reproduction detection method that the mode of applying provides.
The photo is converted into two-value picture in step S102, and the acquisition of the connected domain algorithm based on bianry image is to be checked
The step of key position in the image of survey, can be completed by following steps, be specifically included：
Step S202 carries out color space conversion to input picture, obtains color image.
Specifically, color space conversion is carried out to input picture, obtains color image.For example, can to input picture into
Row YcbCr conversions, obtain the image in the domains YcbCr.
It is possible to further which input picture RGB numerical value is converted according to following formula：
Y=0.2990R+0.5870G+0.1140B
Cb=-0.1787R-0.3313G+0.5000B+128
Cr=0.500R-0.4187G-0.0813B+128.
The color image binaryzation is obtained binary image by step S204.
Specifically, the obtained color images of step S202 are done into binary conversion treatment, binary image can be obtained.For example,
The value of Cb in the image in the domains YcbCr, Cr are subjected to binary conversion treatment, i.e., if 93<Cb<133 and 123<Cr<175, then belong to
The pixel of skin color range.For belonging to the pixel of skin color range, 255 are assigned a value of, is otherwise assigned a value of 0.Therefore two-value can be obtained
Change image I (x, y).
Step S206 carries out vertically mapping and curve smoothing to the binary image, obtains the binary image
In face area.
Specifically, binary image is vertically mapped first, then carries out curve smoothing again, it may be determined that facial regions
Domain.
The face area is carried out gradation conversion, obtains the gray level image of the face area by step S208.
Specifically, according to the face area, the domains the RGB numerical value of face area can be obtained, by the following formula of numerical value
It is converted, gray level image can be obtained.
I=0.259R+0.587G+0.144B.
Step S210 carries out image convolution to the gray level image, and carries out binaryzation to the image after convolution, obtains two
Value convolved image.
Specifically, image convolution is carried out to gray level image using convolution mask, two-value then is carried out to the image after convolution
Change.
Step S212 extracts the key position by connective region search algorithm from the binaryzation convolved image.
Specifically, the connected domain algorithm based on scan line, so that it may to obtain these three key portions of eye, nose and oral area
Position.
The reproduction detection method of the above human face photo obtains eye, nose, oral area these three key areas using the above method
It is higher to obtain efficiency for domain.
Referring to Fig. 3, Fig. 3 is the partial process view of the reproduction detection method provided as one of which embodiment,
In, the step of feature in the extraction key position includes：
Step S302 extracts nose feature using gray level co-occurrence matrixes.
Specifically, for nasal region, gray level co-occurrence matrixes are obtained first, are then selected by gray level co-occurrence matrixes anti-
Difference, at least one of energy and entropy are as nose feature, the characteristics of due to nasal region, using the method for gray level co-occurrence matrixes
Extraction nose feature can obtain better effect.
Specifically, it is assumed that coordinate is any point A of (x, y) and deviates its point B in nose image, and the coordinate of point B is
(x+a, y+b), wherein a, b are according to the preconfigured integer of specific requirements, and the point A and point B constitutes point pair, if institute
The gray value stated a little pair is (f1, f2).The different point in nose image, so that it may to obtain the gray value of difference pair.If institute
The maximum gray scale for stating nose image is L, then shares L2The different gray value of kind.For entire image, each is counted
Then the number that (f1, f2) value occurs is arranged in a square formation, then is normalized to them with (f1, f2) total degree occurred
The probability P (f1, f2) of appearance, resulting matrix are exactly the gray level co-occurrence matrixes.
Specifically, the contrast is also known as contrast, for localized variation in the distribution of the value of metric matrix and image
Amount.The contrast has reacted the clarity of image and the rill depth of texture, and the rill of texture is deeper, and contrast is bigger, and image is got over
Clearly, conversely, rill is more shallow, contrast is smaller, and image is fuzzyyer.The energy refers to square of gray level co-occurrence matrixes each element value
Be evenly distributed degree and texture for measuring the gradation of image of energy response described in the grey scale change degree of stability of image texture
Fineness degree, energy value show that greatly current texture is a kind of texture that variation is relatively stable.The entropy is used for measuring referring to image and including
The randomness of information content, when all values are equal in co-occurrence matrix or when pixel value shows bigger randomness, entropy is bigger,
Therefore entropy shows the complexity of gradation of image distribution, and entropy is bigger, and image is more complicated.
Step S304 extracts eye feature using LBP algorithms.
Specifically, it in ocular, the characteristics of due to ocular, can be obtained more using LBP algorithms extraction eye feature
Good effect.
Specifically, the region of different scale is selected, in such as 3 × 3 square area, using regional center pixel as threshold value,
The gray value of 8 adjacent pixels is compared with it, if surrounding pixel values are more than center pixel value, the position of the pixel
It sets and is marked as 1, be otherwise 0, it is possible thereby to obtain the LBP values of the center pixel, the LBP values can reflect the line in the region
Manage feature.Also referring to Fig. 4, Fig. 4 is the calculating process for the LBP values that an embodiment provides.Described 401 be 3 × 3 just
The gray value of 9 pixels in square region, the wherein gray value of regional center pixel are 5.After described 403 are binaryzation
The square area with gray value, arrange gray value clockwise since the apex angle of upper left, can be obtained one two into
Code 405 processed, described 407 be the LBP values that the binary code is done to the center pixel that the decimal system is converted to：19.
Step 306, oral area feature is extracted using wavelet transformation.
Specifically, according to oral area feature the characteristics of, can preferably be imitated using Wavelet Transformation Algorithm extraction oral area feature
Fruit.
Specifically, it chooses wavelet function according to demand first and sets change of scale, such as Haar small echo letters can be chosen
It counts and sets change of scale as 3.Then to image into every trade convert and rank transformation.Finally the numerical value after transformation is quantified,
It can be obtained feature vector.
It is appreciated that there is no sequencings between above-mentioned steps 302, step 304 and step 306, root can be used
Above three step is combined in any order according to demand.
The reproduction detection method of the above human face photo, obtaining eye, nose, oral area using different methods, these three are crucial
The feature in region further improves detection accuracy.
A kind of reproduction detection device of photo, wherein described device includes：
Key position acquisition module 501, for the photo to be converted to two-value picture, and the connection based on bianry image
Domain algorithm obtains each key position in photo；
Characteristics extraction module 503, for extracting the feature in each key position；
Characteristic value Fusion Module 505 for the characteristic value in each key position to be normalized, and merges each
A normalized characteristic value, obtains fusion feature value；
Classification results acquisition module 507 obtains classification knot for fusion feature value input grader to be carried out classification
Fruit；
Reproduction judgment module 509 confirms that the photo is reproduction if being reproduction class for the classification results.
The reproduction detection device of the above human face photo, is subdivided into key area by face first, then to the key area
Domain carry out feature extraction, to make human face photo reproduction detection method precision higher.
As one embodiment, wherein the key position acquisition module includes：
Face datection unit, for judging in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, being continued to execute by key position acquisition module, and is based on two-value
The connected domain algorithm of image obtains each key position in photo.
As one embodiment, wherein the key position acquisition module includes：
Color image acquiring unit obtains color image for carrying out color space conversion to input picture；
Binary image acquiring unit, for by the color image binaryzation, obtaining binary image；
Face area acquiring unit obtains the binary image for vertically being mapped the binary image
In face area；
Gray level image acquiring unit obtains the ash of the face area for the face area to be carried out gradation conversion
Spend image；
Binaryzation convolved image unit, for the gray level image carry out image convolution, and to the image after convolution into
Row binaryzation obtains binaryzation convolved image；
Key position extraction unit, for by connective region search algorithm, institute to be extracted from the binaryzation convolved image
State key position.
As one embodiment, wherein key position includes nose, eye and oral area；
The characteristics extraction module includes：
Nose feature extraction unit, for extracting nose feature using gray level co-occurrence matrixes；
Eye feature extraction unit, for extracting eye feature using LBP algorithms；
Oral area feature extraction unit, for extracting oral area feature using wavelet transformation.
A kind of computer equipment, including memory, processor and storage can be run on a memory and on a processor
Computer program, wherein when the processor executes described program, realize following steps：
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each pass in photo
Key position；
Extract the feature in each key position；
Characteristic value in each key position is normalized, and merges each normalized characteristic value, is obtained
Obtain fusion feature value；
Fusion feature value input grader is subjected to classification and obtains classification results；
If the classification results are reproduction class, confirm that the photo is reproduction.
The method that computer program in the above computer equipment is realized when being executed by processor, it is first that face is thin
It is divided into key area, the extraction of feature is then carried out to the key area, to makes the reproduction detection method of human face photo
Precision higher.
As one embodiment, wherein the photo is converted to two-value picture by described performed by processor, and is based on
The connected domain algorithm of bianry image obtains each key position in photo, including：
Judge in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, executing, and the connected domain algorithm based on bianry image is obtained and shone
Each key position in piece.
As one embodiment, wherein the photo is converted to two-value picture performed by processor, and is based on two-value
The key position that the connected domain algorithm of image obtains in image to be detected includes：
Color space conversion is carried out to input picture, obtains color image；
By the color image binaryzation, binary image is obtained；
The binary image is vertically mapped, the face area in the binary image is obtained；
The face area is subjected to gradation conversion, obtains the gray level image of the face area；
Image convolution is carried out to the gray level image, and binaryzation is carried out to the image after convolution, obtains binaryzation convolution
Image；
By connective region search algorithm, the key position is extracted from the binaryzation convolved image.
As one embodiment, wherein the key position performed by processor includes nose, eye and oral area；
The step of feature in the extraction key position includes：
Nose feature is extracted using gray level co-occurrence matrixes；
Eye feature is extracted using LBP algorithms；
Oral area feature is extracted using wavelet transformation.
A kind of readable storage medium storing program for executing, the readable storage medium storing program for executing are stored with computer program, and the computer program is located
Reason device realizes following steps when executing：
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each pass in photo
Key position；
Extract the feature in each key position；
Characteristic value in each key position is normalized, and merges each normalized characteristic value, is obtained
Obtain fusion feature value；
Fusion feature value input grader is subjected to classification and obtains classification results；
If the classification results are reproduction class, confirm that the photo is reproduction.
The method that the computer program of the above readable storage medium storing program for executing is realized when being executed by processor, it is first that face is thin
It is divided into key area, the extraction of feature is then carried out to the key area, to makes the reproduction detection method of human face photo
Precision higher.
As one embodiment, wherein can the computer program of storage medium storage realize when being executed by processor
It is described that the photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each key portion in photo
Position, including：
Judge in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, executing, and the connected domain algorithm based on bianry image is obtained and shone
Each key position in piece.
As one embodiment, wherein can the computer program of storage medium storage realize when being executed by processor
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains the crucial portion in image to be detected
Position includes：
Color space conversion is carried out to input picture, obtains color image；
By the color image binaryzation, binary image is obtained；
The binary image is vertically mapped, the face area in the binary image is obtained；
The face area is subjected to gradation conversion, obtains the gray level image of the face area；
Image convolution is carried out to the gray level image, and binaryzation is carried out to the image after convolution, obtains binaryzation convolution
Image；
By connective region search algorithm, the key position is extracted from the binaryzation convolved image.
As one embodiment, wherein can the computer program of storage medium storage realize when being executed by processor
In method, the key position includes nose, eye and oral area；
The step of feature in the extraction key position includes：
Nose feature is extracted using gray level co-occurrence matrixes；
Eye feature is extracted using LBP algorithms；
Oral area feature is extracted using wavelet transformation.
It should be understood by those skilled in the art that, embodiments herein can be provided as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the application
Apply the form of example.Moreover, the application can be used in one or more wherein include computer usable program code computer
The computer program production implemented in usable storage medium (including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)
The form of product.
The application is with reference to method, the flow of equipment (system) and computer program product according to the embodiment of the present application
Figure and/or block diagram describe.It is interpreted as to be realized by computer program instructions each in flowchart and/or the block diagram
The combination of flow and/or box in flow and/or box and flowchart and/or the block diagram.These computer journeys can be provided
Sequence instruct to all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices processor with
Generate a machine so that the instruction generation executed by computer or the processor of other programmable data processing devices is used for
Realize the dress for the function of being specified in one flow of flow chart or multiple flows and/or one box of block diagram or multiple boxes
It sets.
These computer program instructions, which may also be stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that instruction generation stored in the computer readable memory includes referring to
Enable the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one box of block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device so that count
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, in computer or
The instruction executed 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 a box or multiple boxes.
Each technical characteristic of embodiment described above can be combined arbitrarily, to keep description succinct, not to above-mentioned reality
It applies all possible combination of each technical characteristic in example to be all described, as long as however, the combination of these technical characteristics is not deposited
In contradiction, it is all considered to be the range of this specification record.
Several embodiments of the invention above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the inventive concept of the premise, various modifications and improvements can be made, these belong to the protection of the present invention
Range.Therefore, the protection domain of patent of the present invention should be determined by the appended claims.
Claims (10)
1. a kind of reproduction detection method of photo, which is characterized in that the method includes：
The photo is converted into two-value picture, and the connected domain algorithm based on bianry image obtains each key portion in photo
Position；
Extract the characteristic value in each key position；
Characteristic value in each key position is normalized, and merges each normalized characteristic value, is merged
Characteristic value；
Fusion feature value input grader is subjected to classification and obtains classification results；
If the classification results are reproduction class, confirm that the photo is reproduction.
2. reproduction detection method according to claim 1, which is characterized in that described that the photo is converted to binary map
Piece, and the connected domain algorithm based on bianry image obtains each key position in photo, including：
Judge in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, executing, and the connected domain algorithm based on bianry image obtains in photo
Each key position.
3. reproduction detection method according to claim 1 or 2, which is characterized in that the photo is converted into two-value picture,
And the key position that the connected domain algorithm based on bianry image obtains in image to be detected includes：
Color space conversion is carried out to input picture, obtains color image；
By the color image binaryzation, binary image is obtained；
The binary image is vertically mapped, the face area in the binary image is obtained；
The face area is subjected to gradation conversion, obtains the gray level image of the face area；
Image convolution is carried out to the gray level image, and binaryzation is carried out to the image after convolution, obtains binaryzation convolved image；
By connective region search algorithm, the key position is extracted from the binaryzation convolved image.
4. reproduction detection method according to claim 1, which is characterized in that key position includes nose, eye and mouth
Portion；
The step of feature in the extraction key position includes：
Nose feature is extracted using gray level co-occurrence matrixes；
Eye feature is extracted using LBP algorithms；
Oral area feature is extracted using wavelet transformation.
5. a kind of reproduction detection device of photo, which is characterized in that described device includes：
Key position acquisition module, for the photo to be converted to two-value picture, and the connected domain algorithm based on bianry image
Obtain each key position in photo；
Characteristics extraction module, for extracting the characteristic value in each key position；
Characteristic value Fusion Module for the characteristic value in each key position to be normalized, and merges each normalizing
The characteristic value of change obtains fusion feature value；
Classification results acquisition module obtains classification results for fusion feature value input grader to be carried out classification；
Reproduction judgment module confirms that the photo is reproduction if being reproduction class for the classification results.
6. reproduction detection device according to claim 5, which is characterized in that the key position acquisition module includes：
Face datection unit, for judging in the image to be detected whether to include face information；
The photo is converted into two-value picture if so, being continued to execute by key position acquisition module, and is based on bianry image
Connected domain algorithm obtain photo in each key position.
7. reproduction detection device according to claim 5 or 6, which is characterized in that the key position acquisition module includes：
Color image acquiring unit obtains color image for carrying out color space conversion to input picture；
Binary image acquiring unit, for by the color image binaryzation, obtaining binary image；
Face area acquiring unit is obtained for vertically being mapped the binary image in the binary image
Face area；
Gray level image acquiring unit obtains the gray-scale map of the face area for the face area to be carried out gradation conversion
Picture；
Binaryzation convolved image unit for carrying out image convolution to the gray level image, and carries out two to the image after convolution
Value obtains binaryzation convolved image；
Key position extraction unit, for by connective region search algorithm, the pass to be extracted from the binaryzation convolved image
Key position.
8. reproduction detection device according to claim 5, which is characterized in that key position includes nose, eye and mouth
Portion；
The characteristics extraction module includes：
Nose feature extraction unit, for extracting nose feature using gray level co-occurrence matrixes；
Eye feature extraction unit, for extracting eye feature using LBP algorithms；
Oral area feature extraction unit, for extracting oral area feature using wavelet transformation.
9. a kind of computer equipment, including memory, processor and storage are on a memory and the meter that can run on a processor
Calculation machine program, which is characterized in that when the processor executes described program, realize any one of the claim 1-4 sides
The step of method.
10. a kind of readable storage medium storing program for executing, the readable storage medium storing program for executing is stored with computer program, which is characterized in that the calculating
The step of claim 1-4 any one the methods are realized when machine program is executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810196019.6A CN108549836B (en) | 2018-03-09 | 2018-03-09 | Photo copying detection method, device, equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810196019.6A CN108549836B (en) | 2018-03-09 | 2018-03-09 | Photo copying detection method, device, equipment and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108549836A true CN108549836A (en) | 2018-09-18 |
| CN108549836B CN108549836B (en) | 2021-04-06 |
Family
ID=63516018
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810196019.6A Active CN108549836B (en) | 2018-03-09 | 2018-03-09 | Photo copying detection method, device, equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108549836B (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109754059A (en) * | 2018-12-21 | 2019-05-14 | 平安科技（深圳）有限公司 | Reproduction image-recognizing method, device, computer equipment and storage medium |
| CN109859227A (en) * | 2019-01-17 | 2019-06-07 | 平安科技（深圳）有限公司 | Reproduction image detecting method, device, computer equipment and storage medium |
| CN109886309A (en) * | 2019-01-25 | 2019-06-14 | 成都浩天联讯信息技术有限公司 | A method of digital picture identity is forged in identification |
| CN111008651A (en) * | 2019-11-13 | 2020-04-14 | 科大国创软件股份有限公司 | Image reproduction detection method based on multi-feature fusion |
| CN112580621A (en) * | 2020-12-24 | 2021-03-30 | 成都新希望金融信息有限公司 | Identity card copying and identifying method and device, electronic equipment and storage medium |
| CN112950559A (en) * | 2021-02-19 | 2021-06-11 | 山东矩阵软件工程股份有限公司 | Method and device for detecting copied image, electronic equipment and storage medium |
| CN113870454A (en) * | 2021-09-29 | 2021-12-31 | 平安银行股份有限公司 | Attendance checking method and device based on face recognition, electronic equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103116763A (en) * | 2013-01-30 | 2013-05-22 | 宁波大学 | Vivo-face detection method based on HSV (hue, saturation, value) color space statistical characteristics |
| US20140056529A1 (en) * | 2011-03-11 | 2014-02-27 | Omron Corporation | Image processing device and image processing method |
| CN105354856A (en) * | 2015-12-04 | 2016-02-24 | 北京联合大学 | Human matching and positioning method and system based on MSER and ORB |
| CN106650669A (en) * | 2016-12-27 | 2017-05-10 | 重庆邮电大学 | Face recognition method for identifying counterfeit photo deception |
| CN106778704A (en) * | 2017-01-23 | 2017-05-31 | 安徽理工大学 | A kind of recognition of face matching process and semi-automatic face matching system |
| CN106971161A (en) * | 2017-03-27 | 2017-07-21 | 深圳大图科创技术开发有限公司 | Face In vivo detection system based on color and singular value features |
-
2018
- 2018-03-09 CN CN201810196019.6A patent/CN108549836B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140056529A1 (en) * | 2011-03-11 | 2014-02-27 | Omron Corporation | Image processing device and image processing method |
| CN103116763A (en) * | 2013-01-30 | 2013-05-22 | 宁波大学 | Vivo-face detection method based on HSV (hue, saturation, value) color space statistical characteristics |
| CN105354856A (en) * | 2015-12-04 | 2016-02-24 | 北京联合大学 | Human matching and positioning method and system based on MSER and ORB |
| CN106650669A (en) * | 2016-12-27 | 2017-05-10 | 重庆邮电大学 | Face recognition method for identifying counterfeit photo deception |
| CN106778704A (en) * | 2017-01-23 | 2017-05-31 | 安徽理工大学 | A kind of recognition of face matching process and semi-automatic face matching system |
| CN106971161A (en) * | 2017-03-27 | 2017-07-21 | 深圳大图科创技术开发有限公司 | Face In vivo detection system based on color and singular value features |
Non-Patent Citations (1)
| Title |
|---|
| 张恺 等: "一种人眼状态实时检测算法的研究", 《电脑知识与技术》 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109754059A (en) * | 2018-12-21 | 2019-05-14 | 平安科技（深圳）有限公司 | Reproduction image-recognizing method, device, computer equipment and storage medium |
| CN109859227A (en) * | 2019-01-17 | 2019-06-07 | 平安科技（深圳）有限公司 | Reproduction image detecting method, device, computer equipment and storage medium |
| WO2020147398A1 (en) * | 2019-01-17 | 2020-07-23 | 平安科技（深圳）有限公司 | Reproduced image detection method and device, computer device and storage medium |
| CN109886309A (en) * | 2019-01-25 | 2019-06-14 | 成都浩天联讯信息技术有限公司 | A method of digital picture identity is forged in identification |
| CN111008651A (en) * | 2019-11-13 | 2020-04-14 | 科大国创软件股份有限公司 | Image reproduction detection method based on multi-feature fusion |
| CN111008651B (en) * | 2019-11-13 | 2023-04-28 | 科大国创软件股份有限公司 | Image reproduction detection method based on multi-feature fusion |
| CN112580621A (en) * | 2020-12-24 | 2021-03-30 | 成都新希望金融信息有限公司 | Identity card copying and identifying method and device, electronic equipment and storage medium |
| CN112580621B (en) * | 2020-12-24 | 2022-04-29 | 成都新希望金融信息有限公司 | Identity card copying and identifying method and device, electronic equipment and storage medium |
| CN112950559A (en) * | 2021-02-19 | 2021-06-11 | 山东矩阵软件工程股份有限公司 | Method and device for detecting copied image, electronic equipment and storage medium |
| CN113870454A (en) * | 2021-09-29 | 2021-12-31 | 平安银行股份有限公司 | Attendance checking method and device based on face recognition, electronic equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108549836B (en) | 2021-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108549836A (en) | Reproduction detection method, device, equipment and the readable storage medium storing program for executing of photo | |
| Rahmouni et al. | Distinguishing computer graphics from natural images using convolution neural networks | |
| RU2711050C2 (en) | Image and attribute quality, image enhancement and identification of features for identification by vessels and faces and combining information on eye vessels with information on faces and / or parts of faces for biometric systems | |
| CN109657595B (en) | Key feature region matching face recognition method based on stacked hourglass network | |
| CN108764358B (en) | Terahertz image identification method, device and equipment and readable storage medium | |
| CN110175986B (en) | Stereo image visual saliency detection method based on convolutional neural network | |
| CN104915972B (en) | Image processing apparatus, image processing method and program | |
| WO2019169884A1 (en) | Image saliency detection method and device based on depth information | |
| US20120257833A1 (en) | Image processing apparatus, image processing method, and computer readable medium | |
| CN109740572A (en) | A kind of human face in-vivo detection method based on partial color textural characteristics | |
| US8923610B2 (en) | Image processing apparatus, image processing method, and computer readable medium | |
| CN111414856B (en) | Face image generation method and device for realizing user privacy protection | |
| CN111611934A (en) | Face detection model generation and face detection method, device and equipment | |
| CN113128271A (en) | Counterfeit detection of face images | |
| CN111339897B (en) | Living body identification method, living body identification device, computer device, and storage medium | |
| CN113505768A (en) | Model training method, face recognition method, electronic device and storage medium | |
| CN104574358A (en) | Method and apparatus for scene segmentation from focal stack images | |
| CN112836625A (en) | Face living body detection method and device and electronic equipment | |
| CN109886223A (en) | Face identification method, bottom library input method, device and electronic equipment | |
| CN114842524B (en) | Face false distinguishing method based on irregular significant pixel cluster | |
| Purnapatra et al. | Presentation attack detection with advanced cnn models for noncontact-based fingerprint systems | |
| Song et al. | Depth-aware saliency detection using discriminative saliency fusion | |
| US9858296B2 (en) | Representative image selection for image management using face recognition | |
| CN113128428B (en) | Depth map prediction-based in vivo detection method and related equipment | |
| CN111723688B (en) | Human body action recognition result evaluation method and device and electronic equipment |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Luo Jing Inventor after: Chen Shujun Inventor after: Liu Yang Inventor after: Li Hongyan Inventor before: Chen Shujun Inventor before: Liu Yang Inventor before: Li Hongyan |