CN114038030A - Image tampering identification method, device and computer storage medium - Google Patents
Image tampering identification method, device and computer storage medium Download PDFInfo
- Publication number
- CN114038030A CN114038030A CN202111194313.1A CN202111194313A CN114038030A CN 114038030 A CN114038030 A CN 114038030A CN 202111194313 A CN202111194313 A CN 202111194313A CN 114038030 A CN114038030 A CN 114038030A
- Authority
- CN
- China
- Prior art keywords
- image
- target
- detection frame
- external expansion
- target object
- 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
Images
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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Abstract
The application provides an image tampering identification method, equipment and a computer storage medium, which mainly comprise the steps of detecting a target object in an image and determining a target detection frame containing the target object; according to the position information of the target detection frame in the image, carrying out external expansion processing on the target detection frame to obtain an external expansion detection frame of the target object; and identifying the target image in the external expansion detection frame to obtain an identification result that the target object is tampered or not tampered. Therefore, whether the target object in the image is tampered or not can be accurately identified, and the method and the device have stronger generalization capability.
Description
Technical Field
The embodiment of the application relates to the technical field of video identification, in particular to a cloud video method, cloud video equipment and a computer storage medium.
Background
The application of the face recognition technology in life is more and more extensive, and the face recognition technology becomes an important means for personal identity authentication. But along with the convenience and the friendliness of the method, the risk that the identity of the user is impersonated is greatly improved. For example, some lawbreakers can forge the identity of others by using printing paper and masks containing facial information, thereby realizing the attack on the face recognition system.
Currently, the discrimination of such common attack modes is incorporated into face recognition systems. However, as the resolution of smart phone shooting is improved and powerful image processing software is developed, the threshold of image editing and tampering is lower and lower, and the tampering result is more and more real, which also prompts the generation of a new method for forging the identity of other people. Currently, how to distinguish this kind of more real attack mode for forging the identity of others has become a greater challenge for face recognition technology.
The existing discrimination method for human face splicing tampering is mainly divided into two types:
one is a method based on artificial features: the method firstly needs to artificially analyze the face tampered image, and then, a specific processing flow and a characteristic extraction operator are designed in a targeted mode. For example, the image is converted from an RGB space to an HSV space, three primary color plane components are separated, feature extraction of image key points is completed by adopting an SIFT algorithm, and finally, discrimination of a tampered image is completed through feature point matching. The algorithm has the disadvantages that the algorithm is extremely dependent on human experience information, and because the algorithm is a processing method which is designed for the current face tampered image in a targeted manner, the generalization capability is poor when a new face tampered image appears, so that the face tampered attack which may appear in different scenes in an actual face recognition system is difficult to deal with.
The second method is a deep learning-based method: deep learning is one of the latest research trends in the field of artificial intelligence at present, and high-level information of an input image is abstracted by constructing a multi-layer nonlinear feature extraction network, so that the image can be expressed autonomously. However, the current deep learning method for image stitching tamper detection is mostly based on target detection framework improvement, such as dual-stream fast RCNN. Because the method can simultaneously judge and position image splicing tampering, the method comprises a feature extraction network and a candidate region generation network, so that the algorithm architecture is complex, in addition, the input data needs to be simultaneously processed by RGB (red, green and blue) flow and noise flow, and the training and the testing are time-consuming. For the face recognition system, the longer face verification duration greatly reduces the user experience, and meanwhile, the scene with higher real-time requirement is difficult to meet.
In view of the above, a human face tampering identification technology with strong generalization capability and short identification time is needed.
Disclosure of Invention
In view of the foregoing, the present application provides an image tampering identification method, device, and computer storage medium, which have the advantages of strong generalization capability and high processing efficiency.
The present application provides, in a first aspect, an image tampering identification method, including: detecting a target object in an image, and determining a target detection frame containing the target object; according to the position information of the target detection frame in the image, performing external expansion processing on the target detection frame to obtain an external expansion detection frame of the target object; and identifying the target image in the external expansion detection frame to obtain an identification result that the target object is tampered or not tampered.
A second aspect of the present application provides a computer storage medium having stored thereon computer instructions, which, when executed by a processor, cause the processor to perform the method of the first aspect.
The third aspect of the present application provides an image tampering identification device, which includes an image detection module, configured to detect a target object in an image, and determine a target detection frame containing the target object; the image processing module is used for executing external expansion processing aiming at the target detection frame according to the position information of the target detection frame in the image to obtain an external expansion detection frame of the target object; and the image identification module is used for identifying the target image in the outward expansion detection frame and obtaining the identification result that the target object is tampered or not tampered.
To sum up, according to the image tampering identification method, the image tampering identification device, and the computer storage medium provided in the embodiments of the present application, the target detection frame in the target object in the image is obtained, the outward expansion processing is performed on the target detection frame, and then the target image in the outward expansion detection frame is identified, so as to obtain the identification result of whether the target object is tampered, so that the present application can simply judge the face splicing tampered image without performing tampering positioning, and can simplify the processing steps of tampering identification, so that the tampering identification technology of the present application can realize millisecond-level identification feedback, and is suitable for being used in various application scenarios, and can also improve the user experience.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the embodiments of the present application, and other drawings can be obtained by those skilled in the art according to the drawings.
Fig. 1 is a schematic flow chart of an image tampering identification method according to a first embodiment of the present application.
Fig. 2 is a schematic flowchart of an image tampering identification method according to a second embodiment of the present application.
Fig. 3 is a schematic flowchart of an image tampering identification method according to a third embodiment of the present application.
Fig. 4 is a schematic structural diagram of an image detection model according to the present application.
Fig. 5 is a schematic diagram of an internal structure of each inverted residual block in the image detection model of the present application.
Fig. 6 is a schematic structural diagram of an image tamper identification device according to a fifth embodiment of the present application.
Element number
600: an image tampering identification device; 602: an image detection module; 604: an image processing module; 606: and an image identification module.
Detailed Description
In order to make those skilled in the art better understand the technical solutions in the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, but not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application shall fall within the scope of the protection of the embodiments in the present application.
In view of the above-mentioned disadvantages of the prior art, such as poor generalization and long processing time, the present application provides an image tampering identification method, device and computer storage medium, which can solve various technical problems in the prior art, and embodiments of the present application will be described in detail below with reference to the drawings.
First embodiment
Fig. 1 shows a schematic flow chart of an image tampering identification method according to a first embodiment of the present application. As shown in the figure, the present embodiment mainly includes the following processing steps:
step S102, detecting a target object in the image and determining a target detection frame containing the target object.
Optionally, the face object in the image may be detected according to a preset face detection rule, and a target detection frame including the face object may be determined.
And step S104, performing external expansion processing on the target detection frame according to the position information of the target detection frame in the image to obtain an external expansion detection frame of the target object.
Optionally, the position of a corner point of any corner point of the target detection frame in the image and the length and width of the target detection frame may be identified, so as to determine the position information of the target detection frame.
For example, a face detector may perform face detection on an image, acquire the corner position of the upper left corner of the target detection frame in the image and the width and height of the target detection frame, and determine the position information of the target detection frame (i.e., the face object): (x)0,y0,w0,h0) Wherein x is0，y0Coordinates, w, representing the top left corner of the target detection box0，h0Respectively representing width information and height information of the target detection frame.
Optionally, the external expansion processing may be executed for the target detection frame according to the position information of the target detection frame and a preset external expansion multiple, so as to obtain the external expansion detection frame.
In this embodiment, the area ratio between the outward expansion detection frame and the target detection frame satisfies the preset outward expansion multiple, wherein the shortest distance between the target center point of the target detection frame and the left frame of the outward expansion detection frame is equal to the shortest distance between the target center point and the right frame of the outward expansion detection frame, and the shortest distance between the target center point and the upper frame of the outward expansion detection frame is three times the shortest distance between the target center point and the lower frame of the outward expansion detection frame.
Alternatively, the preset flare may be between 1.4-fold and 1.8-fold.
Preferably, the preset flare factor may be set to 1.6 times.
Specifically, as the splicing traces of the face splicing tampered images are more concentrated on the whole head edge part, in order to more fully expose the splicing traces of the tampered images and enable a subsequent feature extraction network to extract more obvious image splicing features, the method and the device perform one-step external expansion processing on the target detection frame containing the face object so as to completely incorporate the head edge part of the face into a subsequent recognition range.
In this embodiment, when the preset expansion factor is set to 1.6 times, the expansion factors of the target detection frame can be set to 0.45, 0.15, 0.3, 0.3 (as shown in fig. 2), and then the position information (x) of the target detection frame is obtained0,y0,w0,h0) And the external expansion multiples of the upper, lower, left and right sides can obtain the position information (x) of the external expansion detection frame1,y1,w1,h1)：
And step S106, identifying the target image in the external expansion detection frame, and obtaining the identification result that the target object is tampered or not tampered.
In this embodiment, the image detection model including the feature extraction network and the classification network may be used to identify the target image in the outward expansion detection frame, so as to obtain the identification result of whether the target object is tampered or not tampered.
In summary, in the image tampering identification method of the embodiment, the target detection including the target object is performed with the outward expansion processing, so as to identify whether the target object in the image is tampered based on the outward expansion detection frame, and by the outward expansion processing means, the edge portion of the target object is fully exposed in the identification area, so that the subsequent image detection model can extract a more obvious image stitching feature, thereby improving the accuracy of the image tampering identification result.
Second embodiment
Fig. 2 shows a process flow diagram of an image tampering identification method according to a second embodiment of the present application. The processing flow of the present embodiment may be executed after step S104 and before step S106. As shown in the figure, the present embodiment mainly includes the following processing steps:
step S202, based on the preset image size, the size of the target image in the external expansion detection frame is adjusted, and an adjusted image of the target image is obtained.
Optionally, the target image including the face object may be acquired according to the position information of the extension detection frame in the image.
Optionally, the target image may be uniformly adjusted to an image size of 128 × 128, but the present invention is not limited thereto, and may also be adjusted to other image sizes, and may be arbitrarily adjusted according to actual requirements.
Step S204, based on preset sharpening rules and gray values corresponding to pixel points in the adjusted image, sharpening processing is executed on the adjusted image, and a sharpened image of the adjusted image is obtained.
Alternatively, the preset sharpening rule may be expressed as:
wherein, the 3 x 3 matrix is a sharpening kernel, IinputRepresenting respective gray-scale values, I, corresponding to respective pixel points in the adjusted imageoutputAnd each enhanced gray value corresponding to each pixel point in the sharpened image is represented.
In summary, the image tampering identification method of the embodiment can help to improve the accuracy of the subsequent image identification result by performing the resizing and sharpening process on the target image in the outward expansion detection frame.
Third embodiment
Fig. 3 is a flowchart illustrating an image tampering identification method according to a third embodiment of the present application. This embodiment mainly shows a specific implementation of the step S106, and as shown in the figure, this embodiment mainly includes the following steps:
step S302, feature extraction is carried out on the target image by utilizing the feature extraction network of the image detection model, and the image features of the target image are obtained.
Alternatively, the target image in the outward expansion detection frame output in step S104 may be directly input into an image detection model for detection, or the target image after being subjected to the size adjustment and sharpening process output in step S204 may be input into the image detection model for detection.
Alternatively, as shown in fig. 4, the feature extraction network of this embodiment may include ten inverted residual blocks connected in sequence, where each inverted residual block may perform feature extraction on the target image in sequence, and output an image feature containing 256 feature dimensions.
Specifically, an original lightweight network (e.g., MobileNet v2) is generally the same as image feature extraction of a general scene, and since texture differences of image features between classes in the general scene are large, structural and texture differences between classes of face data are relatively small, and the data size is far less than that of data of the general scene, there is a high possibility that an overfitting problem is generated in training. In addition, in practical application, models with more parameters often increase inference time, and further influence the use experience of users. By comprehensively considering the two factors, the embodiment compresses the number of Inverted Residual blocks (Inverted Residual blocks) in the original lightweight network (MobileNet v2), reduces the number of original 17 blocks to 10, and reduces the feature dimensionality output by the feature extraction network from original 1280 to 256, so that the network reasoning time can be effectively shortened on the premise of ensuring that the image recognition accuracy is not affected, and the user experience is improved.
In this embodiment, as shown in fig. 5, the internal structure of each inverted residual block is "expanded" by first using a 1 × 1 convolution layer (the convolution layer is followed by a Batch Normalization layer and a ReLu activation layer), that is, the number of channels of the feature map is increased, then, the features are further extracted by using a 3 × 3 convolution layer, then, the feature channels are reduced by using a 1 × 1 convolution layer without ReLu activation to realize compression of the parameters, and finally, the input features and the output of the 1 × 1 convolution are added pixel by pixel to obtain the output.
In addition, the feature extraction network of the present embodiment may adopt an EfficientNet series model in addition to the MobileNet v2 model.
And step S304, carrying out classification and identification on the image characteristics by using a classification network of the image detection model to obtain a classification result that the target object is true or false.
As shown in fig. 4, in the present embodiment, the classification network includes a global average pooling layer, a full connection layer, and a Softmax layer, wherein an input dimension of the full connection layer is a feature dimension, i.e., 256 feature dimensions, output by the feature extraction network.
In step S306, if the classification result of the target object is obtained as true, the identification result that the target object is not tampered is output, and if the classification result that the target object is false is obtained, the identification result that the target object is tampered is output.
In summary, in the embodiment, by improving the original feature extraction network, a more lightweight feature extraction network that only includes 10 inverted residual blocks is designed, which not only can improve the network inference speed, but also can better meet the application requirements of various actual scenes, and improve the user experience.
Fourth embodiment
A fourth embodiment of the present application provides a computer storage medium having computer instructions stored thereon, which, when executed by a processor, cause the processor to perform the method of any one of the first to third embodiments.
Fifth embodiment
Fig. 6 shows an architecture diagram of an image tamper recognition device according to a fifth embodiment of the present application. As shown in the figure, the image tampering identification device 600 of the present embodiment mainly includes an image detection module 602, an image processing module 604, and an image identification module 606.
The image detection module 602 is configured to detect a target object in an image and determine a target detection frame containing the target object.
Optionally, the image detection module 602 further detects a face object in the image according to a preset face detection rule, and determines the target detection frame containing the face object.
The image processing module 604 is configured to perform an extension process on the target detection frame according to the position information of the target detection frame in the image, so as to obtain an extension detection frame of the target object.
Optionally, the image processing module 604 further identifies a corner position of any corner point of the target detection frame in the image, and a length and a width of the target detection frame, and determines the position information of the target detection frame.
Optionally, the image processing module 604 further performs an external expansion process on the target detection frame according to the position information of the target detection frame and a preset external expansion multiple, so as to obtain the external expansion detection frame; the area ratio of the external expansion detection frame to the target detection frame meets the preset external expansion multiple, the shortest distance between the target center point of the target detection frame and the left frame of the external expansion detection frame is equal to the shortest distance between the target center point and the right frame of the external expansion detection frame, and the shortest distance between the target center point and the upper frame of the external expansion detection frame is three times the shortest distance between the target center point and the lower frame of the external expansion detection frame.
Optionally, the preset external expansion multiple is between 1.4 times and 1.8 times, and preferably, the preset external expansion multiple is 1.6 times.
Optionally, the image processing module 604 further adjusts the size of the target image in the outward expansion detection frame based on a preset image size, so as to obtain an adjusted image of the target image.
Optionally, the image processing module 604 further performs a sharpening process on the adjusted image based on a preset sharpening rule and each gray value corresponding to each pixel point in the adjusted image, so as to obtain a sharpened image of the adjusted image;
the preset sharpening rule is expressed as:
wherein the 3 x 3 matrix is a sharpening kernel, the IinputRepresenting the gray values corresponding to the pixel points in the adjusted image, IoutputAnd representing each enhanced gray value corresponding to each pixel point in the sharpened image.
The image recognition module 606 recognizes the target image in the outward expansion detection frame, and obtains a recognition result that the target object is tampered or not tampered.
Optionally, the image recognition module 606 further includes performing feature extraction on the target image by using a feature extraction network of an image detection model to obtain an image feature of the target image; classifying and identifying the image characteristics by using a classification network of the image detection model to obtain a classification result that the target object is true or false; and if the classification result that the target object is true is obtained, outputting the identification result that the target object is not tampered, and if the classification result that the target object is false is obtained, outputting the identification result that the target object is tampered.
Optionally, the feature extraction network includes ten inverted residual blocks connected in sequence, each inverted residual block performs feature extraction on the target image in sequence, and outputs the image feature including 256 feature dimensions.
Optionally, the classification network comprises a global average pooling layer, a fully connected layer, and a Softmax layer.
In addition, the image tampering identification device 600 according to the embodiment of the present invention may also be used to implement other steps in the foregoing image tampering identification method embodiments, and has the beneficial effects of the corresponding method step embodiments, which are not described herein again.
In summary, the image tampering identification method, device and computer storage medium according to the embodiments of the present application first lock a face portion in an image to generate a target detection frame, so that the problems of detail damage and precision loss of the face image due to scaling of an original image can be avoided.
Secondly, according to the method and the device, the edge part of the head can be fully exposed by performing outward expansion, size adjustment and sharpening processing on the target detection frame, so that feature extraction and judgment of a subsequent network are facilitated, and judgment accuracy can be effectively improved.
Moreover, the method and the device finish the classification judgment of the preprocessed face images through a lighter-weight feature extraction network, compared with an original mobilenet v2 and efficientnet series model, the image detection model has less parameter quantity and smaller calculated quantity, the reasoning speed of the model can be effectively improved, the application requirements of various actual scenes are met, and the use experience of users is improved.
Finally, it should be noted that: the above embodiments are only used for illustrating the technical solutions of the embodiments of the present application, and are not limited thereto; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.
Claims (12)
1. An image tampering identification method, comprising:
detecting a target object in an image, and determining a target detection frame containing the target object;
according to the position information of the target detection frame in the image, performing external expansion processing on the target detection frame to obtain an external expansion detection frame of the target object; and
and identifying the target image in the external expansion detection frame to obtain an identification result that the target object is tampered or not tampered.
2. The image tamper recognition method according to claim 1, wherein the detecting a target object in an image and the determining a target detection frame including the target object includes:
and detecting a face object in the image according to a preset face detection rule, and determining the target detection frame containing the face object.
3. The image tamper recognition method according to claim 2, further comprising:
and identifying the corner position of any corner point of the target detection frame in the image and the length and width of the target detection frame, and determining the position information of the target detection frame.
4. The image tampering identification method according to claim 3, wherein the performing, for the target detection frame, an outward expansion process according to the position information of the target detection frame in the image, and obtaining the outward expansion detection frame of the target object includes:
according to the position information of the target detection frame and a preset external expansion multiple, executing external expansion processing on the target detection frame to obtain the external expansion detection frame;
the area ratio of the external expansion detection frame to the target detection frame meets the preset external expansion multiple, the shortest distance between the target center point of the target detection frame and the left frame of the external expansion detection frame is equal to the shortest distance between the target center point and the right frame of the external expansion detection frame, and the shortest distance between the target center point and the upper frame of the external expansion detection frame is three times the shortest distance between the target center point and the lower frame of the external expansion detection frame.
5. The image tampering identification method according to claim 4, wherein the preset external expansion factor is between 1.4 times and 1.8 times, and preferably the preset external expansion factor is 1.6 times.
6. The image tamper recognition method according to claim 4, further comprising:
and adjusting the size of the target image in the external expansion detection frame based on a preset image size to obtain an adjusted image of the target image.
7. The image tamper recognition method according to claim 6, further comprising:
based on a preset sharpening rule and each gray value corresponding to each pixel point in the adjusted image, carrying out sharpening processing on the adjusted image to obtain a sharpened image of the adjusted image;
the preset sharpening rule is expressed as:
wherein the 3 x 3 matrix is a sharpening kernel, the IinputRepresenting the gray values corresponding to the pixel points in the adjusted image, IoutputAnd representing each enhanced gray value corresponding to each pixel point in the sharpened image.
8. The image tampering identification method according to claim 1 or 7, wherein the identifying the target image in the external expansion detection frame, and obtaining the identification result that the target object is tampered or not tampered comprises:
performing feature extraction on the target image by using a feature extraction network of an image detection model to acquire image features of the target image;
classifying and identifying the image characteristics by using a classification network of the image detection model to obtain a classification result that the target object is true or false; and
and if the classification result that the target object is true is obtained, outputting the identification result that the target object is not tampered, and if the classification result that the target object is false is obtained, outputting the identification result that the target object is tampered.
9. The image tamper recognition method according to claim 8, wherein the feature extraction network includes ten inverted residual blocks connected in sequence, each of the inverted residual blocks performing feature extraction on the target image in sequence, and outputting the image feature including 256 feature dimensions.
10. The image tampering identification method according to claim 9, wherein the classification network includes a global averaging pooling layer, a full connection layer, and a Softmax layer.
11. A computer storage medium having stored thereon computer instructions which, when executed by a processor, cause the processor to perform the method of any one of claims 1 to 10.
12. An image tamper recognition device characterized by comprising:
the image detection module is used for detecting a target object in an image and determining a target detection frame containing the target object;
the image processing module is used for executing external expansion processing aiming at the target detection frame according to the position information of the target detection frame in the image to obtain an external expansion detection frame of the target object;
and the image identification module is used for identifying the target image in the outward expansion detection frame and obtaining the identification result that the target object is tampered or not tampered.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111194313.1A CN114038030A (en) | 2021-10-13 | 2021-10-13 | Image tampering identification method, device and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111194313.1A CN114038030A (en) | 2021-10-13 | 2021-10-13 | Image tampering identification method, device and computer storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114038030A true CN114038030A (en) | 2022-02-11 |
Family
ID=80141280
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111194313.1A Pending CN114038030A (en) | 2021-10-13 | 2021-10-13 | Image tampering identification method, device and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114038030A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114943703A (en) * | 2022-05-24 | 2022-08-26 | 闫雪 | Multi-component P map region analysis system |
-
2021
- 2021-10-13 CN CN202111194313.1A patent/CN114038030A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114943703A (en) * | 2022-05-24 | 2022-08-26 | 闫雪 | Multi-component P map region analysis system |
| CN114943703B (en) * | 2022-05-24 | 2023-09-05 | 闫雪 | Multi-component P-map region analysis system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11830230B2 (en) | Living body detection method based on facial recognition, and electronic device and storage medium | |
| CN112381775B (en) | Image tampering detection method, terminal device and storage medium | |
| CN110084135B (en) | Face recognition method, device, computer equipment and storage medium | |
| KR101896357B1 (en) | Method, device and program for detecting an object | |
| KR101781358B1 (en) | Personal Identification System And Method By Face Recognition In Digital Image | |
| CN111275685A (en) | Method, device, equipment and medium for identifying copied image of identity document | |
| CN110838119B (en) | Human face image quality evaluation method, computer device and computer readable storage medium | |
| CN111339897B (en) | Living body identification method, living body identification device, computer device, and storage medium | |
| CN113361495A (en) | Face image similarity calculation method, device, equipment and storage medium | |
| CN110390308B (en) | Video behavior identification method based on space-time confrontation generation network | |
| CN112528902B (en) | Video monitoring dynamic face recognition method and device based on 3D face model | |
| CN111582155B (en) | Living body detection method, living body detection device, computer equipment and storage medium | |
| CN111259792A (en) | Face living body detection method based on DWT-LBP-DCT characteristics | |
| CN114038030A (en) | Image tampering identification method, device and computer storage medium | |
| CN111723614A (en) | Traffic signal lamp identification method and device | |
| JP2007026308A (en) | Image processing method and image processor | |
| CN116229528A (en) | Living body palm vein detection method, device, equipment and storage medium | |
| CN112990047B (en) | Multi-pose face verification method combining face angle information | |
| CN113705307A (en) | Image processing method, device, equipment and storage medium | |
| Rahman et al. | Real-time face-based auto-focus for digital still and cell-phone cameras | |
| JP2010257324A (en) | Image processing apparatus and method, and program | |
| CN115880787A (en) | Face living body detection method, system, terminal equipment and storage medium | |
| CN117392409A (en) | Work clothes identification method, device, equipment and computer readable storage medium | |
| CN117952878A (en) | Image processing method, device, computer equipment and computer readable storage medium | |
| CN114821732A (en) | Living body detection method, model training method, device, electronic device, and medium |
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 |