Deep face counterfeit video detection method based on rPPG signal
A video detection and deep human technology, applied in deception detection, neural learning methods, instruments, etc., can solve problems such as loss of function, poor generalization, complex network structure, etc., to maintain detection performance, good detection performance, and improve robustness sexual effect
Pending Publication Date: 2022-08-09
NANJING UNIV OF SCI & TECH
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Among them, the detection method based on artifacts is that the existing forgery methods will produce some artifacts invisible to the naked eye when generating face forgery videos. Deep learning methods can be used to extract and identify these artifacts, but these artifacts are often The generalization is poor, and it is often only able to detect face forgery videos generated by specific face forgery methods. The method is likely to lose its effect in the future
Based on the data-driven method, the existing image classification network is used for forgery detection. This type of method uses the neural network to automatically find the distinguishing features between the real and the forged, and can achieve good detection performance. However, the network used by this type of method The structure is relatively complex, so it takes a lot of time and space resources
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0031] The present invention will be further illustrated below in conjunction with specific examples, which are implemented on the premise of the technical solutions of the present invention. It should be understood that these examples are only used to illustrate the present invention and not to limit the scope of the present invention.
[0032] like figure 1 As shown, the rPPG signal-based deep face forgery video detection method of the present application includes the following steps:
[0033] Step 1: Get the face sequence from the video;
[0034] Input the video to be detected into the face extraction module, and the face extraction module uses the FaceMesh method to extract the face and the corresponding face landmark points from each frame of the video, and generate a face sequence of 128 frames. The acquired face image is attached figure 2 shown.
[0035] Step 2: Extract the rPPG signal from selecting a specific face region of interest and using a green single-channe...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a deep face counterfeit video detection method based on rPPG signals, and belongs to the field of artificial intelligence safety. Comprises: acquiring a face sequence from a video; selecting a specific face region of interest and extracting rPPG signals by adopting a method based on a green single channel; and performing forgery identification according to the acquired rPPG signal by using a CNN-based classifier. According to the invention, based on the characteristic that most of the current counterfeiting methods are difficult to simulate the heart rate signal of the human body, the rPPG signal of the specific part of the video face is extracted to carry out the detection of the counterfeited face, so that the face counterfeited videos generated by various counterfeiting methods can be effectively detected; and meanwhile, the detection performance of the method on the high-compression forged video is improved by adopting an rPPG signal extraction method with certain video compression resistance.
Description
technical field [0001] The invention belongs to the technical field of artificial intelligence security, and in particular relates to a deep face forgery video detection method based on an rPPG signal. Background technique [0002] In recent years, with the continuous development of artificial intelligence technology, face forgery technology has also made great progress. At present, this technology has been widely used in various cultural and entertainment industries, greatly enriching daily life. But what is worrying is that this technology brings some benefits to society but also brings great risks to society as a whole. Therefore, how to effectively detect the fake face videos generated by this kind of technology has become an urgent problem to be solved. [0003] Current deep face forgery detection techniques are mainly divided into three categories: artifact-based detection methods, data-driven detection methods, and information inconsistency-based detection methods. ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06V20/40G06V40/16G06V40/40G06V10/25G06V10/82G06N3/04G06N3/08G06K9/00
CPCG06V20/41G06V40/168G06V40/161G06V10/25G06V40/40G06V40/172G06V10/82G06N3/08G06N3/045G06F2218/02G06F2218/04G06F2218/12
Inventor 张伟耿嘉仪练智超
Owner NANJING UNIV OF SCI & TECH