Two-dimensional depth false graph generalization detection method, system and device and storage medium

A deep fake image and detection method technology, applied in the field of fake image detection, can solve problems such as weak effect and weak generalization of fake image detectors, and achieve the effects of improving detection performance, improving detection accuracy, and strong generalization effect

Pending Publication Date: 2022-04-29
NORTHWEST UNIV(CN)
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Aiming at the problems in the prior art that the generalization of detectors is weak and the effect is weak in actual application scenarios, the present invention provides a generalization detection method, system, device and storage medium for two-dimensional deep false images, which utilize various The common upsampling feature of the three-dimensional fake image generation method uses frequency domain conversion and neural network technology to solve the problem of weak generalization of the fake image detector and weak effect in actual application scenarios

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Two-dimensional depth false graph generalization detection method, system and device and storage medium
  • Two-dimensional depth false graph generalization detection method, system and device and storage medium
  • Two-dimensional depth false graph generalization detection method, system and device and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0060] For the two-dimensional deep fake image generalization detection method, the following steps are included:

[0061] Step 1, use the fake image and real image generated by ProGAN technology as the training data set;

[0062] Step 2: Perform post-processing simulation operations on the collected fake images and real images, and perform frequency-domain preprocessing to obtain preprocessed frequency-domain images. The original renderings without frequency-domain preprocessing are as follows: figure 2 shown;

[0063] Among them, the frequency domain preprocessing method is as follows:

[0064] Step 21, extracting images in the training set with a probability of 10% to perform JPEG image compression processing to obtain post-processing simulation images, such as image 3 shown;

[0065] Step 22, extracting images in the training set with a probability of 10% to perform Gaussian blur processing to obtain post-processing simulated images, such as Figure 4 shown;

[0066...

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

No PUM Login to view more

Abstract

The invention relates to the related technical field of false image detection, provides a two-dimensional depth false image generalization detection method, system and device, and a storage medium, and firstly overcomes the difficulty that an image is difficult to carry out generalization detection by using a unique detector due to a large difference between a DeepFake technology and a GAN network technology in the prior art. Through the detection method, the limitation of a current detection method based on a spatial domain is overcome, frequency domain information is extracted for detection, and the detection performance is improved. According to the method, common up-sampling features of various different two-dimensional false image generation methods are utilized, and the problems that a false image detector is weak in generalization and poor in effect in an actual application scene are solved by using frequency domain conversion and neural network technologies.

Description

technical field [0001] The invention relates to the related technical field of fake image detection, in particular to a generalized detection method, system, device and storage medium of a two-dimensional deep fake image. Background technique [0002] With the continuous development of neural network technology, various methods of generating fake images have been born and gradually improved. [0003] This type of generation technology can complete many different image processing tasks such as face replacement, fake face generation, and attribute editing. The birth of two-dimensional fake image generation technology has seriously threatened the reputation of people in the modern information society, the security of network communication, and the reliability of media reports. In contrast, fake image detection technology emerged as the times require. At present, the airspace detection model for a single generation method of fake images has achieved excellent results, but most...

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
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/42G06V10/774G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/214
Inventor 刘晓宁蒋艺江雨林芃樾冯龙
Owner NORTHWEST UNIV(CN)
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap