CN107657259A - Distorted image detection method, electronic installation and readable storage medium storing program for executing - Google Patents
Distorted image detection method, electronic installation and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN107657259A CN107657259A CN201710916508.XA CN201710916508A CN107657259A CN 107657259 A CN107657259 A CN 107657259A CN 201710916508 A CN201710916508 A CN 201710916508A CN 107657259 A CN107657259 A CN 107657259A
- Authority
- CN
- China
- Prior art keywords
- image
- detected
- block
- distorted
- data set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
Abstract
The present invention relates to a kind of distorted image detection method, electronic installation and readable storage medium storing program for executing, this method includes：Block segmentation is carried out to image to be detected, described image to be detected is divided into the grid image block of some default specifications, and the tampering detection feature of predetermined number is extracted from some grid image blocks；Described image to be detected is identified for tampering detection feature and predetermined identification model based on predetermined number, and analyzes whether described image to be detected is tampered according to recognition result；Wherein, the predetermined identification model is to be trained obtained deep neural network model beforehand through to being labeled with the different samples pictures for distorting indication character.The invention enables that can detect the image type distorted to be not limited in jpeg format, realize and different type distorted image is detected.
Description
Technical field
The present invention relates to field of computer technology, more particularly to a kind of distorted image detection method, electronic installation and readable
Storage medium.
Background technology
It is widely available with the digital video editing instrument such as Adobe Photoshop, ACDSee, make more and more users
Digital photos are carried out with freely random modification, while the malicious user opportunity of illegal objective is also carried to some, not
Illegal operation is carried out to picture material in the case of authorized, such as synthesis is false, edits in violation of rules and regulations, so as to damage.It is existing
Traditional algorithm can only be directed to image type and distort and detected for the picture of jpeg format, enabling carry out distorted image inspection
The picture format of survey is relatively simple.
The content of the invention
It is an object of the invention to provide a kind of distorted image detection method, electronic installation and readable storage medium storing program for executing, it is intended to
The enabled image type distorted that detects is not limited in JPEG, realizes and distorting for different type image is detected.
To achieve the above object, the present invention provides a kind of electronic installation, and the electronic installation includes memory, processor,
The distorted image detecting system that can be run on the processor, described image tampering detection system are stored with the memory
Following steps are realized during by the computing device：
A, block segmentation is carried out to image to be detected, described image to be detected is divided into the grid image of some default specifications
Block, and extract from some grid image blocks the tampering detection feature of predetermined number；
B, the tampering detection feature based on predetermined number and predetermined identification model are carried out to described image to be detected
Identification, and analyze whether described image to be detected is tampered according to recognition result；Wherein, the predetermined identification model is
Obtained deep neural network model is trained beforehand through to being labeled with the different samples pictures for distorting indication character.
Preferably, when described image tampering detection system realizes the step A by the computing device, including：
By described image to be detected from primitive color space reflection to YCrCb color spaces；
Block segmentation is carried out to described image to be detected, described image to be detected is divided into some 16*16 grid image block；
Using the western wavelet decomposition fortune of the more shellfishes of two dimension in each YCrCb color space components on each grid image block
Calculate, obtain corresponding pixel coefficient mapping, to each pixel coefficient mapping calculation collect statistics coefficient, and apply more Bei Xi
Orthogonal Daubechies Orthogonal wavelets D2-D5 become to the collect statistics coefficient of all grid image blocks
Change, obtain 450 features as tampering detection feature.
Preferably, the deep neural network model that it is all full articulamentum that the predetermined identification model, which is, the depth
The number of the neuron of neural network model is 450,500,256,128,2 respectively, wherein, 450 be the number of input feature vector,
500th, 256,128 be default hidden feature number, 2 be final class categories number.
Preferably, the training process of the predetermined identification model is as follows：
C, the image pattern of predetermined number is obtained；
D, block segmentation is carried out to each image pattern, each image pattern is divided into 16*16 grid image block, and to every
One grid image block is labeled, and the grid image block distorted is labeled as into 1, and the grid image block do not distorted is labeled as 0, with
To each image pattern distort the mark of indication character；
E, by all image patterns according to X：Y ratio is divided into the first data set and the second data set, in the first data set
Image pattern quantity be more than the second data set in image pattern quantity, the first data set is as training set, the second data set
As test set, wherein, X is more than 0, Y and is more than 0；
F, the predetermined identification model is trained using each image pattern in the first data set；
G, the accuracy rate of the identification model of training is verified using each image pattern in the second data set, if accuracy rate is big
In or equal to default accuracy rate, then train and terminate, or, if accuracy rate is less than default accuracy rate, increase image pattern
Quantity simultaneously re-executes above-mentioned steps D, E, F and G.
In addition, to achieve the above object, the present invention also provides a kind of distorted image detection method, described image tampering detection
Method includes：
Step 1: carrying out block segmentation to image to be detected, described image to be detected is divided into the grid of some default specifications
Image block, and extract from some grid image blocks the tampering detection feature of predetermined number；
Step 2: tampering detection feature and predetermined identification model based on predetermined number are to described image to be detected
It is identified, and analyzes whether described image to be detected is tampered according to recognition result；Wherein, the predetermined identification mould
Type is to be trained obtained deep neural network model beforehand through to being labeled with the different samples pictures for distorting indication character.
Preferably, the step 1 includes：
By described image to be detected from primitive color space reflection to YCrCb color spaces；
Block segmentation is carried out to described image to be detected, described image to be detected is divided into some 16*16 grid image block；
Using the western wavelet decomposition fortune of the more shellfishes of two dimension in each YCrCb color space components on each grid image block
Calculate, obtain corresponding pixel coefficient mapping, to each pixel coefficient mapping calculation collect statistics coefficient, and apply more Bei Xi
Orthogonal Daubechies Orthogonal wavelets D2-D5 become to the collect statistics coefficient of all grid image blocks
Change, obtain 450 features as tampering detection feature.
Preferably, the deep neural network model that it is all full articulamentum that the predetermined identification model, which is, the depth
The number of the neuron of neural network model is 450,500,256,128,2 respectively, wherein, 450 be the number of input feature vector,
500th, 256,128 be default hidden feature number, 2 be final class categories number.
Preferably, the training process of the predetermined identification model is as follows：
C, the image pattern of predetermined number is obtained；
D, block segmentation is carried out to each image pattern, each image pattern is divided into 16*16 grid image block, and to every
One grid image block is labeled, and the grid image block distorted is labeled as into 1, and the grid image block do not distorted is labeled as 0, with
To each image pattern distort the mark of indication character；
E, by all image patterns according to X：Y ratio is divided into the first data set and the second data set, in the first data set
Image pattern quantity be more than the second data set in image pattern quantity, the first data set is as training set, the second data set
As test set, wherein, X is more than 0, Y and is more than 0；
F, the predetermined identification model is trained using each image pattern in the first data set；
G, the accuracy rate of the identification model of training is verified using each image pattern in the second data set, if accuracy rate is big
In or equal to default accuracy rate, then train and terminate, or, if accuracy rate is less than default accuracy rate, increase image pattern
Quantity simultaneously re-executes above-mentioned steps D, E, F and G.
Preferably, it is described to analyze that the step of whether described image to be detected is tampered includes according to recognition result：
If identifying that described image to be detected has using the predetermined identification model distorts indication character, divide
Described image to be detected is analysed to be tampered；
If identifying that described image to be detected does not have using the predetermined identification model distorts indication character,
Described image to be detected is analyzed to be not tampered with.
Further, to achieve the above object, the present invention also provides a kind of computer-readable recording medium, the computer
Readable storage medium storing program for executing is stored with distorted image detecting system, and described image tampering detection system can be held by least one processor
OK, so that the step of at least one computing device distorted image detection method described above.
Distorted image detection method, system and readable storage medium storing program for executing proposed by the present invention, by being usurped based on being labeled with difference
Change deep neural network model that some samples pictures of indication character are trained to obtain image to be detected to be identified,
And analyze whether described image to be detected is tampered according to recognition result.Image to be detected progress block is divided into by then passing through
Some grid image blocks, tampering detection feature is extracted from each grid image block, and utilize the good depth god of training in advance
The tampering detection feature of extraction is identified through network model, to identify whether image to be detected distorts vestige, not
Dependent on traditional JPEG compression trace detection algorithm so that the image type distorted can be detected and be not limited in jpeg format,
Realization detects to different type distorted image.
Brief description of the drawings
Fig. 1 is the running environment schematic diagram of the preferred embodiment of distorted image detecting system 10 of the present invention；
Fig. 2 is the structural representation of each one optional identification model of embodiment of the present invention；
Fig. 3 is the schematic flow sheet of the embodiment of distorted image detection method one of the present invention.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, it is right below in conjunction with drawings and Examples
The present invention is further elaborated.It should be appreciated that specific embodiment described herein is only to explain the present invention, not
For limiting the present invention.Based on the embodiment in the present invention, those of ordinary skill in the art are not before creative work is made
The every other embodiment obtained is put, belongs to the scope of protection of the invention.
It should be noted that the description for being related to " first ", " second " etc. in the present invention is only used for describing purpose, and can not
It is interpreted as indicating or implies its relative importance or imply the quantity of the technical characteristic indicated by indicating.Thus, define " the
One ", at least one this feature can be expressed or be implicitly included to the feature of " second ".In addition, the skill between each embodiment
Art scheme can be combined with each other, but must can be implemented as basis with those of ordinary skill in the art, when technical scheme
With reference to occurring conflicting or will be understood that the combination of this technical scheme is not present when can not realize, also not in application claims
Protection domain within.
The present invention provides a kind of distorted image detecting system.Referring to Fig. 1, be distorted image detecting system 10 of the present invention compared with
The running environment schematic diagram of good embodiment.
In the present embodiment, described distorted image detecting system 10 is installed and run in electronic installation 1.The electronics fills
Putting 1 may include, but be not limited only to, memory 11, processor 12 and display 13.Fig. 1 illustrate only the electricity with component 11-13
Sub-device 1, it should be understood that being not required for implementing all components shown, the implementation that can be substituted is more or less
Component.
The memory 11 comprises at least a type of readable storage medium storing program for executing, and the memory 11 is in certain embodiments
Can be the internal storage unit of the electronic installation 1, such as the hard disk or internal memory of the electronic installation 1.The memory 11 exists
It in other embodiments can also be the External memory equipment of the electronic installation 1, such as be equipped with the electronic installation 1 slotting
Connect formula hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash memory
Block (Flash Card) etc..Further, the memory 11 can also both include the internal storage unit of the electronic installation 1
Also External memory equipment is included.The memory 11, which is used to store, is installed on the application software of the electronic installation 1 and all kinds of numbers
According to, such as program code of described image tampering detection system 10 etc..The memory 11 can be also used for temporarily storing
Data through exporting or will export.
The processor 12 can be in certain embodiments a central processing unit (Central Processing Unit,
CPU), microprocessor or other data processing chips, for running the program code stored in the memory 11 or processing number
According to, such as perform described image tampering detection system 10 etc..
The display 13 can be light-emitting diode display, liquid crystal display, touch-control liquid crystal display in certain embodiments
And OLED (Organic Light-Emitting Diode, Organic Light Emitting Diode) touches device etc..The display 13 is used
In being shown in the information that is handled in the electronic installation 1 and for showing visual user interface, such as image to be detected,
Distorted image information etc..The part 11-13 of the electronic installation 1 is in communication with each other by system bus.
Described image tampering detection system 10 includes at least one computer-readable finger being stored in the memory 11
Order, at least one computer-readable instruction can be performed by the processor 12, to realize each embodiment of the application.
Wherein, following steps are realized when above-mentioned distorted image detecting system 10 is performed by the processor 12：
Step S1, block segmentation is carried out to image to be detected, described image to be detected is divided into the grid of some default specifications
Image block, and extract from some grid image blocks the tampering detection feature of predetermined number.
In the present embodiment, distorted image detecting system 10 receives the inspection of the distorted image comprising picture to be detected that user sends
Request is surveyed, the picture to be detected includes but is not limited to the various Format Types such as JPEG, PNG, GIF.For example, receiving user passes through hand
The distorted image detection request that the terminals such as machine, tablet personal computer, self-help terminal equipment are sent, user is such as received in mobile phone, flat board electricity
The distorted image detection request sent in the terminals such as brain, self-help terminal equipment in preassembled client, or receive user
The distorted image detection request sent on browser in the terminals such as mobile phone, tablet personal computer, self-help terminal equipment.
Block segmentation is carried out to the image to be detected received, described image to be detected is divided into some default specification (such as N*
M, N and M are positive integer) grid image block, and extract from some grid image blocks the tampering detection feature of predetermined number.
For example, in a kind of optional embodiment, first by described image to be detected from primitive color space (for example, RGB color is empty
Between) it is transformed into YCrCb color spaces；Block segmentation is carried out to described image to be detected, described image to be detected is divided into some 16*
16 grid image block.Default small echo letter is applied in each YCrCb color space components on each grid image block again
Two-dimentional more western small echos of shellfish (2D Daubechies Wavelet) of number such as multiple ranks (for example, 3 ranks) are decomposed, to obtain
Pixel coefficient mapping (Coefficient Map) corresponding to multiple (for example, 30), to each pixel coefficient mapping calculation
Collect statistics coefficient is (for example, pixel criterion corresponding to each grid image block pixel coefficient mapping is poor, pixel average and pixel are total
With and horizontal direction, vertical direction and diagonally opposed coefficient), and the application for example more shellfishes of wavelet function western orthogonal Daubechies
Orthogonal wavelets D2-D5 obtain multiple collect statistics coefficients on each grid image block, and to all grids
The collect statistics coefficient of image block enters line translation, obtains multiple collect statistics coefficients as tampering detection feature.It is for example, available
450 features are as tampering detection feature.
In the present embodiment, it is divided into the grid image block of 16*16 pixels to carry out feature described image to be detected progress block
Extraction, compared to 32*32 pixels are broken generally into traditional dividing processing, grid image block is reduced into tradition in the present embodiment divides
Cut a quarter of specification, each minutia of the finer described image to be detected of extraction of energy, moreover, western small using more shellfishes
Ripple, the more shellfishes western mode such as orthogonal obtains the collect statistics coefficient of the grid image block of every 16*16 pixels as tampering detection
Feature, it can make it that the tampering detection feature of extraction is more accurate.
Step S2, tampering detection feature and predetermined identification model based on predetermined number are to described image to be detected
It is identified, and analyzes whether described image to be detected is tampered according to recognition result；Wherein, the predetermined identification mould
Type is to be trained obtained deep neural network model beforehand through to being labeled with the different samples pictures for distorting indication character.
From image to be detected block segmentation after some grid image blocks in extract predetermined number tampering detection feature it
Afterwards, described image to be detected can be carried out based on the predetermined number tampering detection feature and predetermined identification model extracted
Identification, and analyze whether described image to be detected is tampered according to recognition result.The identification model can be beforehand through to a large amount of marks
It is marked with the different samples pictures for distorting indication character to be identified constantly be trained, learn, verifying, optimizing, by it
The model that various differences distort indication character can be recognized accurately by being trained to.For example, can be that common picture distorts indication character
Some samples pictures corresponding to preparation, such as picture ad-hoc location (such as the upper left corner, the lower left corner, the upper right corner, the lower right corner, top,
It is micro- end, just medium) in stickup signature, watermark etc. distort picture, can mark corresponding to difference distort indication character, and for mark
It is marked with the different samples pictures for distorting indication character to be trained, learn, verify, optimize, to generate identification model.The identification
Model can use depth convolutional neural networks model (Convolutional Neural Network, CNN) model, stack automatic
Encoder etc., is not limited herein.
In a kind of optional embodiment, the training process of the identification model is as follows：
C, the image pattern of predetermined number (for example, 100,000) is obtained；
D, block segmentation is carried out to each image pattern, each image pattern is divided into 16*16 grid image block, and to every
One grid image block is labeled, and the grid image block distorted is labeled as into 1, and the grid image block do not distorted is labeled as 0, with
To each image pattern distort the mark of indication character.In the present embodiment, what is finally marked is not usurping for image pattern
Change part, but image pattern distorts vestige, is the equal of to distort the feature of vestige in training, rather than distort the spy of content
Sign, moreover, the vestige distorted is labeled as into 1 (positive class), other to be labeled as 0 (negative class), the grid image block of segmentation is reduced into biography
The a quarter (32*32 pixels are changed to 16*16 pixels) of system segmentation specification, preferably can be trained and identify in practical application often
See it is elongated distort vestige, practicality is stronger.
E, by all image patterns according to X：Y is (for example, 8:2) ratio is divided into the first data set and the second data set, the
Image pattern quantity in one data set is more than the image pattern quantity in the second data set, the first data set as training set,
Second data set as test set, wherein, X be more than 0, Y be more than 0；
F, the predetermined identification model is trained using each image pattern in the first data set；
G, the accuracy rate of the identification model of training is verified using each image pattern in the second data set, if accuracy rate is big
In or equal to default accuracy rate (for example, 95%), then train and terminate, or, if accuracy rate is less than default accuracy rate, increase
The quantity of image pattern simultaneously re-executes above-mentioned steps D, E, F and G, until the accuracy rate of the identification model of training is more than or equal to
Default accuracy rate.
Described image to be detected is identified using the identification model trained, and according to being analyzed recognition result
When whether image to be detected is tampered, if identifying that described image to be detected has using the identification model distorts vestige spy
Sign, then analyze described image to be detected and be tampered；If identified using the predetermined identification model described to be detected
Image is then analyzed described image to be detected and is not tampered with without indication character is distorted.
Compared with prior art, the present embodiment is by being entered based on being labeled with the different some samples pictures for distorting indication character
Row trains obtained deep neural network model that image to be detected is identified, and described to be checked according to recognition result analysis
Whether altimetric image is tampered.Image to be detected progress block is divided into some grid image blocks by then passing through, from each grid
Tampering detection feature, and the tampering detection using the good deep neural network model of training in advance to extraction are extracted in image block
Feature is identified, and to identify whether image to be detected distorts vestige, is not rely on traditional JPEG compression vestige inspection
Method of determining and calculating so that the image type distorted can be detected and be not limited in jpeg format, realize and carried out to different type distorted image
Detection.
In an optional embodiment, on the basis of above-mentioned Fig. 1 embodiment, the predetermined identification model is
All be the deep neural network model of full articulamentum, the number of the neuron of the deep neural network model is 450 respectively, 500,
256th, 128,2, wherein, 450 be the number of input feature vector, and 500,256,128 be the number of default hidden feature, and 2 be final
The number of class categories.
It is the structural representation of each one optional identification model of embodiment of the present invention refering to Fig. 2.From image to be detected block
Predetermined number such as 450 tampering detection features, and extract 450 are usurped are extracted in some grid image blocks after segmentation
Change detection feature to input to the input layer of the deep neural network model as the input feature vector of deep neural network model, then according to
The secondary hidden layer 1 through deep neural network model, hidden layer 2, hidden layer 3 carry out the processing such as Feature Dimension Reduction filtering, reach final
Classification layer, export two kinds of classification results finally identifying, be " not distorting " and " having distorted " in the present embodiment so that
Complete to detect the distorted image of image to be detected.
As shown in figure 3, Fig. 3 is the schematic flow sheet of the embodiment of distorted image detection method one of the present invention, the distorted image
Detection method comprises the following steps：
Step S10, block segmentation is carried out to image to be detected, described image to be detected is divided into the grid of some default specifications
Image block, and extract from some grid image blocks the tampering detection feature of predetermined number.
In the present embodiment, distorted image detecting system receives the detection of the distorted image comprising picture to be detected that user sends
Request, the picture to be detected include but is not limited to the various Format Types such as JPEG, PNG, GIF.For example, receiving user passes through hand
The distorted image detection request that the terminals such as machine, tablet personal computer, self-help terminal equipment are sent, user is such as received in mobile phone, flat board electricity
The distorted image detection request sent in the terminals such as brain, self-help terminal equipment in preassembled client, or receive user
The distorted image detection request sent on browser in the terminals such as mobile phone, tablet personal computer, self-help terminal equipment.
Block segmentation is carried out to the image to be detected received, described image to be detected is divided into some default specification (such as N*
M, N and M are positive integer) grid image block, and extract from some grid image blocks the tampering detection feature of predetermined number.
For example, in a kind of optional embodiment, first by described image to be detected from primitive color space (for example, RGB color is empty
Between) it is transformed into YCrCb color spaces；Block segmentation is carried out to described image to be detected, described image to be detected is divided into some 16*
16 grid image block.Default small echo letter is applied in each YCrCb color space components on each grid image block again
Two-dimentional more western small echos of shellfish (2D Daubechies Wavelet) of number such as multiple ranks (for example, 3 ranks) are decomposed, to obtain
Pixel coefficient mapping (Coefficient Map) corresponding to multiple (for example, 30), to each pixel coefficient mapping calculation
Collect statistics coefficient is (for example, pixel criterion corresponding to each grid image block pixel coefficient mapping is poor, pixel average and pixel are total
With and horizontal direction, vertical direction and diagonally opposed coefficient), and the application for example more shellfishes of wavelet function western orthogonal Daubechies
Orthogonal wavelets D2-D5 obtain multiple collect statistics coefficients on each grid image block, and to all grids
The collect statistics coefficient of image block enters line translation, obtains multiple collect statistics coefficients as tampering detection feature.It is for example, available
450 features are as tampering detection feature.
In the present embodiment, it is divided into the grid image block of 16*16 pixels to carry out feature described image to be detected progress block
Extraction, compared to 32*32 pixels are broken generally into traditional dividing processing, grid image block is reduced into tradition in the present embodiment divides
Cut a quarter of specification, each minutia of the finer described image to be detected of extraction of energy, moreover, western small using more shellfishes
Ripple, the more shellfishes western mode such as orthogonal obtains the collect statistics coefficient of the grid image block of every 16*16 pixels as tampering detection
Feature, it can make it that the tampering detection feature of extraction is more accurate.
Step S20, tampering detection feature and predetermined identification model based on predetermined number are to the mapping to be checked
Analyze whether described image to be detected is tampered as being identified, and according to recognition result；Wherein, the predetermined identification
Model is to be trained obtained deep neural network mould beforehand through to being labeled with the different samples pictures for distorting indication character
Type.
From image to be detected block segmentation after some grid image blocks in extract predetermined number tampering detection feature it
Afterwards, described image to be detected can be carried out based on the predetermined number tampering detection feature and predetermined identification model extracted
Identification, and analyze whether described image to be detected is tampered according to recognition result.The identification model can be beforehand through to a large amount of marks
It is marked with the different samples pictures for distorting indication character to be identified constantly be trained, learn, verifying, optimizing, by it
The model that various differences distort indication character can be recognized accurately by being trained to.For example, can be that common picture distorts indication character
Some samples pictures corresponding to preparation, such as picture ad-hoc location (such as the upper left corner, the lower left corner, the upper right corner, the lower right corner, top,
It is micro- end, just medium) in stickup signature, watermark etc. distort picture, can mark corresponding to difference distort indication character, and for mark
It is marked with the different samples pictures for distorting indication character to be trained, learn, verify, optimize, to generate identification model.The identification
Model can use depth convolutional neural networks model (Convolutional Neural Network, CNN) model, stack automatic
Encoder etc., is not limited herein.
In a kind of optional embodiment, the training process of the identification model is as follows：
C, the image pattern of predetermined number (for example, 100,000) is obtained；
D, block segmentation is carried out to each image pattern, each image pattern is divided into 16*16 grid image block, and to every
One grid image block is labeled, and the grid image block distorted is labeled as into 1, and the grid image block do not distorted is labeled as 0, with
To each image pattern distort the mark of indication character.In the present embodiment, what is finally marked is not usurping for image pattern
Change part, but image pattern distorts vestige, is the equal of to distort the feature of vestige in training, rather than distort the spy of content
Sign, moreover, the vestige distorted is labeled as into 1 (positive class), other to be labeled as 0 (negative class), the grid image block of segmentation is reduced into biography
The a quarter (32*32 pixels are changed to 16*16 pixels) of system segmentation specification, preferably can be trained and identify in practical application often
See it is elongated distort vestige, practicality is stronger.
E, by all image patterns according to X：Y is (for example, 8:2) ratio is divided into the first data set and the second data set, the
Image pattern quantity in one data set is more than the image pattern quantity in the second data set, the first data set as training set,
Second data set as test set, wherein, X be more than 0, Y be more than 0；
F, the predetermined identification model is trained using each image pattern in the first data set；
G, the accuracy rate of the identification model of training is verified using each image pattern in the second data set, if accuracy rate is big
In or equal to default accuracy rate (for example, 95%), then train and terminate, or, if accuracy rate is less than default accuracy rate, increase
The quantity of image pattern simultaneously re-executes above-mentioned steps D, E, F and G, until the accuracy rate of the identification model of training is more than or equal to
Default accuracy rate.
Described image to be detected is identified using the identification model trained, and according to being analyzed recognition result
When whether image to be detected is tampered, if identifying that described image to be detected has using the identification model distorts vestige spy
Sign, then analyze described image to be detected and be tampered；If identified using the predetermined identification model described to be detected
Image is then analyzed described image to be detected and is not tampered with without indication character is distorted.
Compared with prior art, the present embodiment is by being entered based on being labeled with the different some samples pictures for distorting indication character
Row trains obtained deep neural network model that image to be detected is identified, and described to be checked according to recognition result analysis
Whether altimetric image is tampered.Image to be detected progress block is divided into some grid image blocks by then passing through, from each grid
Tampering detection feature, and the tampering detection using the good deep neural network model of training in advance to extraction are extracted in image block
Feature is identified, and to identify whether image to be detected distorts vestige, is not rely on traditional JPEG compression vestige inspection
Method of determining and calculating so that the image type distorted can be detected and be not limited in jpeg format, realize and carried out to different type distorted image
Detection.
In an optional embodiment, on the basis of above-described embodiment, the predetermined identification model is all to be
The deep neural network model of full articulamentum, the number of the neuron of the deep neural network model is 450 respectively, 500,256,
128th, 2, wherein, 450 be the number of input feature vector, and 500,256,128 be the number of default hidden feature, and 2 be final classification
The number of classification.
It is the structural representation of each one optional identification model of embodiment of the present invention refering to Fig. 2.From image to be detected block
Predetermined number such as 450 tampering detection features, and extract 450 are usurped are extracted in some grid image blocks after segmentation
Change detection feature to input to the input layer of the deep neural network model as the input feature vector of deep neural network model, then according to
The secondary hidden layer 1 through deep neural network model, hidden layer 2, hidden layer 3 carry out the processing such as Feature Dimension Reduction filtering, reach final
Classification layer, export two kinds of classification results finally identifying, be " not distorting " and " having distorted " in the present embodiment so that
Complete to detect the distorted image of image to be detected.
In addition, the present invention also provides a kind of computer-readable recording medium, the computer-readable recording medium storage has
Distorted image detecting system, described image tampering detection system can be by least one computing devices, so that described at least one
The step of distorted image detection method in computing device such as above-mentioned embodiment, the step S10 of the distorted image detection method,
The specific implementation process such as S20 as described above, will not be repeated here.
It should be noted that herein, term " comprising ", "comprising" or its any other variant are intended to non-row
His property includes, so that process, method, article or device including a series of elements not only include those key elements, and
And also include the other element being not expressly set out, or also include for this process, method, article or device institute inherently
Key element.In the absence of more restrictions, the key element limited by sentence "including a ...", it is not excluded that including this
Other identical element also be present in the process of key element, method, article or device.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can add the mode of required general hardware platform to realize by software, naturally it is also possible to realized by hardware, but a lot
In the case of the former be more preferably embodiment.Based on such understanding, technical scheme is substantially in other words to existing
The part that technology contributes can be embodied in the form of software product, and the computer software product is stored in a storage
In medium (such as ROM/RAM, magnetic disc, CD), including some instructions to cause a station terminal equipment (can be mobile phone, calculate
Machine, server, air conditioner, or network equipment etc.) perform method described in each embodiment of the present invention.
Above by reference to the preferred embodiments of the present invention have been illustrated, not thereby limit to the interest field of the present invention.On
State that sequence number of the embodiment of the present invention is for illustration only, do not represent the quality of embodiment.Patrolled in addition, though showing in flow charts
Order is collected, but in some cases, can be with the step shown or described by being performed different from order herein.
Those skilled in the art do not depart from the scope of the present invention and essence, can have a variety of flexible programs to realize the present invention,
It can be used for another embodiment for example as the feature of one embodiment and obtain another embodiment.All technologies with the present invention
The all any modification, equivalent and improvement made within design, all should be within the interest field of the present invention.
Claims (10)
1. a kind of electronic installation, it is characterised in that the electronic installation includes memory, processor, is stored on the memory
There is the distorted image detecting system that can be run on the processor, described image tampering detection system is by the computing device
Shi Shixian following steps：
A, block segmentation is carried out to image to be detected, described image to be detected is divided into the grid image block of some default specifications, and
The tampering detection feature of predetermined number is extracted from some grid image blocks；
B, described image to be detected is identified for the tampering detection feature based on predetermined number and predetermined identification model,
And analyze whether described image to be detected is tampered according to recognition result；Wherein, the predetermined identification model is advance
By being trained obtained deep neural network model to being labeled with the different samples pictures for distorting indication character.
2. electronic installation as claimed in claim 1, it is characterised in that described image tampering detection system is held by the processor
When row realizes the step A, including：
By described image to be detected from primitive color space reflection to YCrCb color spaces；
Block segmentation is carried out to described image to be detected, described image to be detected is divided into some 16*16 grid image block；
Using the more western wavelet decomposition computings of shellfish of two dimension in each YCrCb color space components on each grid image block, obtain
Mapped to corresponding pixel coefficient, to each pixel coefficient mapping calculation collect statistics coefficient, and the more shellfishes of application are western orthogonal
Daubechies Orthogonal wavelets D2-D5 enter line translation to the collect statistics coefficient of all grid image blocks, obtain
To 450 features as tampering detection feature.
3. electronic installation as claimed in claim 2, it is characterised in that it is all to connect entirely that the predetermined identification model, which is,
The deep neural network model of layer, the number of the neuron of the deep neural network model is 450,500,256,128,2 respectively,
Wherein, 450 be input feature vector number, 500,256,128 be the number of default hidden feature, and 2 is final class categories
Number.
4. electronic installation as claimed in claim 1 or 2, it is characterised in that the training of the predetermined identification model
Journey is as follows：
C, the image pattern of predetermined number is obtained；
D, block segmentation is carried out to each image pattern, each image pattern is divided into 16*16 grid image block, and to each net
Table images block is labeled, and the grid image block distorted is labeled as into 1, the grid image block do not distorted is labeled as 0, with to each
Individual image pattern distort the mark of indication character；
E, by all image patterns according to X：Y ratio is divided into the first data set and the second data set, the figure in the first data set
As sample size be more than the second data set in image pattern quantity, the first data set is as training set, the second data set conduct
Test set, wherein, X is more than 0, Y and is more than 0；
F, the predetermined identification model is trained using each image pattern in the first data set；
G, using in the second data set each image pattern verify training identification model accuracy rate, if accuracy rate be more than or
Person is equal to default accuracy rate, then training terminates, or, if accuracy rate is less than default accuracy rate, increase the quantity of image pattern
And re-execute above-mentioned steps D, E, F and G.
5. a kind of distorted image detection method, it is characterised in that described image altering detecting method includes：
Step 1: carrying out block segmentation to image to be detected, described image to be detected is divided into the grid image of some default specifications
Block, and extract from some grid image blocks the tampering detection feature of predetermined number；
Step 2: tampering detection feature and predetermined identification model based on predetermined number are carried out to described image to be detected
Identification, and analyze whether described image to be detected is tampered according to recognition result；Wherein, the predetermined identification model is
Obtained deep neural network model is trained beforehand through to being labeled with the different samples pictures for distorting indication character.
6. distorted image detection method as claimed in claim 5, it is characterised in that the step 1 includes：
By described image to be detected from primitive color space reflection to YCrCb color spaces；
Block segmentation is carried out to described image to be detected, described image to be detected is divided into some 16*16 grid image block；
Using the more western wavelet decomposition computings of shellfish of two dimension in each YCrCb color space components on each grid image block, obtain
Mapped to corresponding pixel coefficient, to each pixel coefficient mapping calculation collect statistics coefficient, and the more shellfishes of application are western orthogonal
Daubechies Orthogonal wavelets D2-D5 enter line translation to the collect statistics coefficient of all grid image blocks, obtain
To 450 features as tampering detection feature.
7. distorted image detection method as claimed in claim 6, it is characterised in that the predetermined identification model is all
The deep neural network model of full articulamentum, the number of the neuron of the deep neural network model is 450 respectively, 500,
256th, 128,2, wherein, 450 be the number of input feature vector, and 500,256,128 be the number of default hidden feature, and 2 be final
The number of class categories.
8. the distorted image detection method as described in claim 5 or 6, it is characterised in that the predetermined identification model
Training process it is as follows：
C, the image pattern of predetermined number is obtained；
D, block segmentation is carried out to each image pattern, each image pattern is divided into 16*16 grid image block, and to each net
Table images block is labeled, and the grid image block distorted is labeled as into 1, the grid image block do not distorted is labeled as 0, with to each
Individual image pattern distort the mark of indication character；
E, by all image patterns according to X：Y ratio is divided into the first data set and the second data set, the figure in the first data set
As sample size be more than the second data set in image pattern quantity, the first data set is as training set, the second data set conduct
Test set, wherein, X is more than 0, Y and is more than 0；
F, the predetermined identification model is trained using each image pattern in the first data set；
G, using in the second data set each image pattern verify training identification model accuracy rate, if accuracy rate be more than or
Person is equal to default accuracy rate, then training terminates, or, if accuracy rate is less than default accuracy rate, increase the quantity of image pattern
And re-execute above-mentioned steps D, E, F and G.
9. the distorted image detection method as described in claim 5 or 6, it is characterised in that described that institute is analyzed according to recognition result
Stating the step of whether image to be detected is tampered includes：
If identifying that described image to be detected has using the predetermined identification model distorts indication character, institute is analyzed
Image to be detected is stated to be tampered；
If identify that described image to be detected without indication character is distorted, is analyzed using the predetermined identification model
Described image to be detected is not tampered with.
10. a kind of computer-readable recording medium, it is characterised in that be stored with image on the computer-readable recording medium and usurp
Change detecting system, realized when described image tampering detection system is executed by processor as any one of claim 5 to 9
The step of distorted image detection method.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710916508.XA CN107657259A (en) | 2017-09-30 | 2017-09-30 | Distorted image detection method, electronic installation and readable storage medium storing program for executing |
| PCT/CN2017/108765 WO2019061661A1 (en) | 2017-09-30 | 2017-10-31 | Image tamper detecting method, electronic device and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710916508.XA CN107657259A (en) | 2017-09-30 | 2017-09-30 | Distorted image detection method, electronic installation and readable storage medium storing program for executing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107657259A true CN107657259A (en) | 2018-02-02 |
Family
ID=61117280
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710916508.XA Pending CN107657259A (en) | 2017-09-30 | 2017-09-30 | Distorted image detection method, electronic installation and readable storage medium storing program for executing |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN107657259A (en) |
| WO (1) | WO2019061661A1 (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108230227A (en) * | 2018-02-06 | 2018-06-29 | 中国银行股份有限公司 | A kind of recognition methods of distorted image, device and electronic equipment |
| CN108510483A (en) * | 2018-03-23 | 2018-09-07 | 云南师范大学 | A kind of calculating using VLAD codings and SVM generates color image tamper detection method |
| CN109344747A (en) * | 2018-09-17 | 2019-02-15 | 平安科技（深圳）有限公司 | A kind of recognition methods that distorting figure, storage medium and server |
| CN109657600A (en) * | 2018-12-14 | 2019-04-19 | 广东工业大学 | A kind of video area removes altering detecting method and device |
| CN110619631A (en) * | 2019-09-16 | 2019-12-27 | 中山大学 | Super-resolution image detection method based on residual error network |
| CN110895811A (en) * | 2019-11-05 | 2020-03-20 | 泰康保险集团股份有限公司 | Image tampering detection method and device |
| CN110942456A (en) * | 2019-11-25 | 2020-03-31 | 深圳前海微众银行股份有限公司 | Tampered image detection method, device, equipment and storage medium |
| CN111131322A (en) * | 2019-12-31 | 2020-05-08 | 奇安信科技集团股份有限公司 | Network behavior detection method and device, computer equipment and storage medium |
| CN111368342A (en) * | 2020-03-13 | 2020-07-03 | 众安信息技术服务有限公司 | Image tampering identification model training method, image tampering identification method and device |
| CN111882525A (en) * | 2020-07-01 | 2020-11-03 | 上海品览数据科技有限公司 | Image reproduction detection method based on LBP watermark characteristics and fine-grained identification |
| CN112150483A (en) * | 2020-09-27 | 2020-12-29 | 深圳壹账通智能科技有限公司 | Picture tampering detection method and device, terminal device and storage medium |
| CN112150450A (en) * | 2020-09-29 | 2020-12-29 | 武汉大学 | Image tampering detection method and device based on dual-channel U-Net model |
| US20210110010A1 (en) * | 2019-10-10 | 2021-04-15 | Baidu Usa Llc | Method and system for signing an artificial intelligence watermark using a query |
| CN112839141A (en) * | 2019-11-25 | 2021-05-25 | 佳能株式会社 | Image processing system, image processing method, and storage medium |
| CN112950564A (en) * | 2021-02-23 | 2021-06-11 | 北京三快在线科技有限公司 | Image detection method and device, storage medium and electronic equipment |
| CN113450330A (en) * | 2021-06-29 | 2021-09-28 | 山东大学 | Image copying-pasting tampering detection method based on segmentation and depth convolution network |
| CN114612798A (en) * | 2022-03-09 | 2022-06-10 | 云南大学 | Satellite image tampering detection method based on Flow model |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111583180B (en) * | 2020-04-03 | 2022-12-20 | 苏宁云计算有限公司 | Image tampering identification method and device, computer equipment and storage medium |
| CN111582116B (en) * | 2020-04-29 | 2022-09-13 | 腾讯科技（深圳）有限公司 | Video erasing trace detection method, device, equipment and storage medium |
| CN113674195A (en) * | 2020-05-13 | 2021-11-19 | 中国移动通信集团有限公司 | Image detection method, device, equipment and storage medium |
| CN112508039B (en) * | 2020-12-08 | 2024-04-02 | 中国银联股份有限公司 | Image detection method and device |
| CN112907598B (en) * | 2021-02-08 | 2023-03-10 | 东南数字经济发展研究院 | Method for detecting falsification of document and certificate images based on attention CNN |
| CN113065592A (en) * | 2021-03-31 | 2021-07-02 | 上海商汤智能科技有限公司 | Image classification method and device, electronic equipment and storage medium |
| CN114764858B (en) * | 2022-06-15 | 2022-11-01 | 深圳大学 | Copy-paste image identification method and device, computer equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101916442A (en) * | 2010-08-05 | 2010-12-15 | 大连理工大学 | Method for robustly positioning tampered region by utilizing GLCM characteristic |
| CN105989330A (en) * | 2015-02-03 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Picture detection method and apparatus |
| CN106558058A (en) * | 2016-11-29 | 2017-04-05 | 北京图森未来科技有限公司 | Parted pattern training method, lane segmentation method, control method for vehicle and device |
| CN106846303A (en) * | 2016-12-30 | 2017-06-13 | 平安科技（深圳）有限公司 | Distorted image detection method and device |
| US20170200274A1 (en) * | 2014-05-23 | 2017-07-13 | Watrix Technology | Human-Shape Image Segmentation Method |
| CN106951899A (en) * | 2017-02-24 | 2017-07-14 | 李刚毅 | Method for detecting abnormality based on image recognition |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102073978B (en) * | 2010-12-15 | 2013-01-23 | 北京交通大学 | Method and system for identifying and recovering digital images by utilizing irregular region segmentation |
| CN103544692A (en) * | 2012-07-13 | 2014-01-29 | 深圳市智信达软件有限公司 | Blind detection method for tamper with double-compressed JPEG (joint photographic experts group) images on basis of statistical judgment |
| CN105427350B (en) * | 2015-12-28 | 2018-12-07 | 辽宁师范大学 | Color reproduction image altering detecting method based on local quaternary number Uniformly bounded |
| US9881209B2 (en) * | 2016-03-28 | 2018-01-30 | Wipro Limited | Methods and systems for detecting tampering in a document image |
-
2017
- 2017-09-30 CN CN201710916508.XA patent/CN107657259A/en active Pending
- 2017-10-31 WO PCT/CN2017/108765 patent/WO2019061661A1/en active Application Filing
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101916442A (en) * | 2010-08-05 | 2010-12-15 | 大连理工大学 | Method for robustly positioning tampered region by utilizing GLCM characteristic |
| US20170200274A1 (en) * | 2014-05-23 | 2017-07-13 | Watrix Technology | Human-Shape Image Segmentation Method |
| CN105989330A (en) * | 2015-02-03 | 2016-10-05 | 阿里巴巴集团控股有限公司 | Picture detection method and apparatus |
| CN106558058A (en) * | 2016-11-29 | 2017-04-05 | 北京图森未来科技有限公司 | Parted pattern training method, lane segmentation method, control method for vehicle and device |
| CN106846303A (en) * | 2016-12-30 | 2017-06-13 | 平安科技（深圳）有限公司 | Distorted image detection method and device |
| CN106951899A (en) * | 2017-02-24 | 2017-07-14 | 李刚毅 | Method for detecting abnormality based on image recognition |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108230227A (en) * | 2018-02-06 | 2018-06-29 | 中国银行股份有限公司 | A kind of recognition methods of distorted image, device and electronic equipment |
| CN108510483A (en) * | 2018-03-23 | 2018-09-07 | 云南师范大学 | A kind of calculating using VLAD codings and SVM generates color image tamper detection method |
| CN108510483B (en) * | 2018-03-23 | 2020-10-20 | 云南师范大学 | Method for generating color image tampering detection by adopting VLAD coding and SVM calculation |
| CN109344747A (en) * | 2018-09-17 | 2019-02-15 | 平安科技（深圳）有限公司 | A kind of recognition methods that distorting figure, storage medium and server |
| CN109344747B (en) * | 2018-09-17 | 2024-01-05 | 平安科技（深圳）有限公司 | Tamper graph identification method, storage medium and server |
| CN109657600B (en) * | 2018-12-14 | 2024-02-27 | 广东工业大学 | Video area removal tampering detection method and device |
| CN109657600A (en) * | 2018-12-14 | 2019-04-19 | 广东工业大学 | A kind of video area removes altering detecting method and device |
| CN110619631A (en) * | 2019-09-16 | 2019-12-27 | 中山大学 | Super-resolution image detection method based on residual error network |
| US11704390B2 (en) * | 2019-10-10 | 2023-07-18 | Baidu Usa Llc | Method and system for signing an artificial intelligence watermark using a query |
| US20210110010A1 (en) * | 2019-10-10 | 2021-04-15 | Baidu Usa Llc | Method and system for signing an artificial intelligence watermark using a query |
| CN110895811A (en) * | 2019-11-05 | 2020-03-20 | 泰康保险集团股份有限公司 | Image tampering detection method and device |
| CN110942456A (en) * | 2019-11-25 | 2020-03-31 | 深圳前海微众银行股份有限公司 | Tampered image detection method, device, equipment and storage medium |
| CN110942456B (en) * | 2019-11-25 | 2024-01-23 | 深圳前海微众银行股份有限公司 | Tamper image detection method, device, equipment and storage medium |
| CN112839141B (en) * | 2019-11-25 | 2024-01-09 | 佳能株式会社 | Image processing system, image processing method, and storage medium |
| US11797857B2 (en) | 2019-11-25 | 2023-10-24 | Canon Kabushiki Kaisha | Image processing system, image processing method, and storage medium |
| CN112839141A (en) * | 2019-11-25 | 2021-05-25 | 佳能株式会社 | Image processing system, image processing method, and storage medium |
| CN111131322A (en) * | 2019-12-31 | 2020-05-08 | 奇安信科技集团股份有限公司 | Network behavior detection method and device, computer equipment and storage medium |
| CN111368342B (en) * | 2020-03-13 | 2022-07-05 | 众安信息技术服务有限公司 | Image tampering identification model training method, image tampering identification method and device |
| CN111368342A (en) * | 2020-03-13 | 2020-07-03 | 众安信息技术服务有限公司 | Image tampering identification model training method, image tampering identification method and device |
| CN111882525A (en) * | 2020-07-01 | 2020-11-03 | 上海品览数据科技有限公司 | Image reproduction detection method based on LBP watermark characteristics and fine-grained identification |
| CN112150483A (en) * | 2020-09-27 | 2020-12-29 | 深圳壹账通智能科技有限公司 | Picture tampering detection method and device, terminal device and storage medium |
| CN112150450B (en) * | 2020-09-29 | 2021-03-16 | 武汉大学 | Image tampering detection method and device based on dual-channel U-Net model |
| CN112150450A (en) * | 2020-09-29 | 2020-12-29 | 武汉大学 | Image tampering detection method and device based on dual-channel U-Net model |
| CN112950564B (en) * | 2021-02-23 | 2022-04-01 | 北京三快在线科技有限公司 | Image detection method and device, storage medium and electronic equipment |
| CN112950564A (en) * | 2021-02-23 | 2021-06-11 | 北京三快在线科技有限公司 | Image detection method and device, storage medium and electronic equipment |
| CN113450330A (en) * | 2021-06-29 | 2021-09-28 | 山东大学 | Image copying-pasting tampering detection method based on segmentation and depth convolution network |
| CN113450330B (en) * | 2021-06-29 | 2022-03-18 | 山东大学 | Image copying-pasting tampering detection method based on segmentation and depth convolution network |
| CN114612798A (en) * | 2022-03-09 | 2022-06-10 | 云南大学 | Satellite image tampering detection method based on Flow model |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2019061661A1 (en) | 2019-04-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107657259A (en) | Distorted image detection method, electronic installation and readable storage medium storing program for executing | |
| CN107862303B (en) | Information identifying method, electronic device and the readable storage medium storing program for executing of form class diagram picture | |
| TWI665639B (en) | Method and device for detecting tampering of images | |
| CN107194398B (en) | Vehicle damages recognition methods and the system at position | |
| CN104063712B (en) | A kind of information of vehicles extracting method and system | |
| CN105426530B (en) | Trademark retrieving method, device and system | |
| CN105957059B (en) | Electronic component missing part detection method and system | |
| CN105574550A (en) | Vehicle identification method and device | |
| CN109740424A (en) | Traffic violations recognition methods and Related product | |
| CN106228129A (en) | A kind of human face in-vivo detection method based on MATV feature | |
| CN101930549B (en) | Second generation curvelet transform-based static human detection method | |
| CN110488368B (en) | Contraband identification method and device based on dual-energy X-ray security inspection machine | |
| CN107622489A (en) | A kind of distorted image detection method and device | |
| CN111898538A (en) | Certificate authentication method and device, electronic equipment and storage medium | |
| CN110245544A (en) | A kind of method and device of determining dead ship condition | |
| CN110457996B (en) | Video moving object tampering evidence obtaining method based on VGG-11 convolutional neural network | |
| CN110490150A (en) | A kind of automatic auditing system of picture violating the regulations and method based on vehicle retrieval | |
| CN114881698A (en) | Advertisement compliance auditing method and device, electronic equipment and storage medium | |
| CN103310235A (en) | Steganalysis method based on parameter identification and estimation | |
| CN113887438A (en) | Watermark detection method, device, equipment and medium for face image | |
| CN110533704A (en) | Fake method, device, equipment and medium are tested in the identification of ink label | |
| CN110889366A (en) | Method and system for judging user interest degree based on facial expression | |
| CN110443306B (en) | Authenticity identification method for wine cork | |
| CN110222660B (en) | Signature authentication method and system based on dynamic and static feature fusion | |
| CN114677333A (en) | Image contrast enhancement detection method based on histogram |
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 | ||
| REG | Reference to a national code |
Ref country code: HK Ref legal event code: DE Ref document number: 1247373 Country of ref document: HK |
|
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180202 |
|
| RJ01 | Rejection of invention patent application after publication |