Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Anti-jpeg compression forgery image detection method

An image detection and image technology, applied in image enhancement, image analysis, image data processing and other directions, can solve problems such as loss of performance, and achieve the effect of robust performance

Active Publication Date: 2021-11-30
INST OF AUTOMATION CHINESE ACAD OF SCI
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The current deepfake detection model will lose a lot of performance in the face of JPEG compression, so it is a very important issue to effectively resist the impact of JPEG compression during the detection process

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
  • Anti-jpeg compression forgery image detection method
  • Anti-jpeg compression forgery image detection method
  • Anti-jpeg compression forgery image detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0084] 1 Experimental setup

[0085] Diverse Fake Face Dataset (DFFD) and FaceForensics ++ (FF++) are used as the data sets for this experiment. The Diverse Fake Face Dataset dataset consists of multiple publicly available sub-datasets generated by open-source code. Real pictures and fake pictures with different resolutions and image quality are obtained through various ways. Faceforensics++ is a forensic dataset consisting of 1000 original video sequences, which contains five face forgery methods, namely: Deepfakes, Face2Face, FaceSwap, NeuralTextures and FaceShifter. These data are selected from Youtube videos. All videos have continuous and unoccluded faces, which can enable the generative model to successfully generate fake faces. At the same time, the face binary mask information is provided in the dataset, so the dataset can be used for classification or segmentation tasks.

[0086] These two data sets were selected as the experimental data and reclassified according ...

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 provides a counterfeit image detection method against JPEG compression, comprising: intercepting the face area in the original image, deleting useless background information, and then adjusting the obtained face image to a fixed size to obtain the face image text; Described face image text is converted into YCbCr color space component by RGB color space component, obtains YCbCr image text; Described YCbCr image text is divided into a series of 8 * 8 pixel blocks; Each 8 * 8 in the described YCbCr image text The component data of each color space channel of the block of pixels is carried out discrete cosine transform, and YCbCr color space component is converted into 192 frequency channels, and described YCbCr image text is transformed into the data of 192 frequency channels after DCT transformation; In 192 Select the middle and low frequency channel data from the data of three frequency channels; input the middle and low frequency channel data into the CNN network for image detection.

Description

technical field [0001] The invention relates to the field of image detection, in particular to a forged image detection method against JPEG compression. Background technique [0002] With the continuous development of forgery technology, the technology of forgery detection has also been improved rapidly. From the perspective of detection, it can be divided into five categories: detection based on physiological characteristics, detection based on motion patterns, detection based on pixel artifacts, detection based on frequency domain and detection based on GAN fingerprints. [0003] 1. Detection based on physiological characteristics [0004] The detection based on physiological characteristics mainly refers to starting from the perspective of human physiological information, because although the fake video is generated with high quality, it lacks human physiological information. In the early forged videos, there was a lack of data of human eyes closing. Based on this clue,...

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 Patents(China)
IPC IPC(8): G06K9/00G06N3/04G06N3/08G06T7/90
CPCG06N3/08G06T7/90G06T2207/30201G06T2207/20081G06T2207/20084G06T2207/10024G06V40/172G06V40/161G06N3/045
Inventor 董晶王伟彭勃王建文项伟樊红兴
Owner INST OF AUTOMATION CHINESE ACAD OF SCI
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com