A face anti-counterfeiting method based on local motion patterns

Abstract

The invention relates to a face anti-fake method based on a local motion pattern, comprising the following steps: (1) detecting a face image area collected by a camera, and positioning face key points; (2) collecting motion information of a face area and a non-face area in a local area where the face key points are located; (3) calculating a local motion pattern of the face according to the obtained local motion information at all the key points; and (4) using a pre-configured pattern classifier to judge whether the face is fake based on the local motion pattern of the face. The beneficial effect is that the face anti-fake method can be effectively combined with a practical face recognition system to quickly and effectively distinguish a true face from a fake face with little need for user interaction.

Description

technical field [0001] The invention relates to the fields of computer vision and pattern recognition, and the research on the face anti-counterfeiting method in the field of biological feature recognition, in particular to a face anti-counterfeiting method based on local motion patterns. Background technique [0002] At present, biometric identification technology has been widely used in all aspects of daily life. Face biometric recognition technology, due to its advantages of convenience, user-friendliness, and non-contact, has achieved rapid development in recent years. These developments are reflected in various research fields, including face detection, face feature Extraction, classifier design, hardware device fabrication, etc. However, face-based biometric identification still faces some challenges at the application level, the most prominent of which is the security of the identification system; as a device for identification, they are easily counterfeited by a cri...

AI Technical Summary

Problems solved by technology

[0004] 1. Face anti-counterfeiting method based on skin reflection characteristics: Starting from the reflection characteristics of human face skin, some researchers use multi-spectral acquisition methods for face anti-counterfeiting; The characteristic of different reflectance achieves the purpose of face anti-counterfeiting; the research content of this type of method is to find a suitable spectrum to make the difference between true and false face skin the largest; however, this type of method has the following obvious deficiencies: 1 ) is only tested on a very small amount of data, so the performance cannot be fully evaluated; 2) the selected spectral band cannot be sensed by commonly used cameras, and special photosensitive devices need to be deployed, which increases hardware overhead; 3) additional photosensitive devices require Development of targeted signal conversion circuits, increasing compatibility issues with existing systems

[0005] 2. Face anti-counterfeiting method based on texture difference: The face anti-counterfeiting method based on micro-texture has the following assumptions: compared with the real face collected by the same device, there are details missing or different, and these details are The difference causes the difference in the microtexture of the image; this assumption is true in most cases, the fake face is made by using the real face picture, taking the printed photo as an example, the illegal user first prints the photo On the paper, and then put the printed face photo in front of the face recognition system to attack; in this process, there will be at least two links that cause differences, one is the printing link, it is impossible for the printer to reproduce the photo without distortion The second is the secondary imaging of printed photos. It is impossible for the acquisition equipment to perfectly capture the content of the photos; in addition, the difference in surface shape between the real face and the printed face, and the difference in local highlights cause the difference in microtexture between the two

[0006] 3. Face anti-counterfeiting method based on motion: This method aims to determine whether the collected face is a real face by detecting the physiological response of the face; considering that the real face has more autonomy than the fake face , this type of method requires the user to perform a specified action as the basis for judgment; commonly used interaction methods include blinking, shaking the head, mouth movements, etc.; in addition to detection methods based on local motion, there is another type of method based on the entire head The reason why this kind of method is effective is that there are obvious differences in the three-dimensional structure of the photo and the face, which makes the acquired head movement patterns also have certain differences; in order to further improve the anti-counterfeiting performance of the face, a method based on multiple A modal face anti-counterfeiting method is proposed; this method requires the user to read the specified text content, and then judges the authenticity of the face by analyzing whether the user's lip movements match the corresponding voice content; however, this method based on human-computer interaction The anti-counterfeiting method requires the user to perform specific actions, and the user experience is too high, which makes the user experience poor. At the same time, the long authentication time is also a major drawback of the above method.

[0007] Among the above three methods, the face anti-counterfeiting method based on motion has the advantages of not being affected by lighting conditions and image quality. However, when this type of method extracts motion features, it does not accurately locate each area of ​​the face, so it cannot Accurately describe the actual motion state of the collected face; for example, some methods roughly divide the collected image into a rectangular area of ​​the face and a background area, and judge the authenticity of the face by comparing the motion state of the two. However, the rectangular frame The determined face area contains a large number of background areas, so that the real face is likely to be misjudged as a fake face; at the same time, in this case, the fake face can easily deceive people by folding and distorting Face anti-counterfeiting system; therefore, how to accurately locate the face area and non-face area, and find the most discriminative local area to extract the highly discriminative local motion pattern information is whether the face anti-counterfeiting system can be applied in practice The essential

Images