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

A detection method for image tampered by splicing

A detection method and image technology, applied in the field of image analysis, can solve the problems of single and incomplete image tampering features, inability to realize end-to-end pixel-level positioning, etc., and achieve the effect of improving accuracy and robustness

Inactive Publication Date: 2021-08-03
HEBEI UNIV OF TECH
View PDF16 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] The technical problem to be solved by the present invention is to provide a detection method of spliced ​​and falsified images, which is a detection method of spliced ​​and falsified images based on light source maps and pyramid feature maps, using two convolutional neural networks with the same structure respectively Extract the multi-stage features of the spliced ​​falsified image and its corresponding light source map, combine the multi-scale information, fuse and up-sample the two sets of multi-stage features to obtain the pyramid feature map, and pass the different layers of the pyramid feature map through the region generation network to obtain To tamper with the candidate area, generate a fixed-size feature map through ROI Align, classify the fixed-size feature map, perform bounding box regression and mask prediction, and finally obtain the bounding box and pixel-level positioning of the tampered area, and complete the detection of the stitched tampered image , overcomes the defect that the image tampering features extracted by the existing technology are single and incomplete, it is easy to ignore the tampering target with a small area, and it cannot achieve end-to-end pixel-level positioning

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
  • A detection method for image tampered by splicing
  • A detection method for image tampered by splicing
  • A detection method for image tampered by splicing

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0072] The first step is to generate the corresponding light source map from the input image:

[0073] First, the input image is divided into similar regions, namely superpixels, and then the light source color is estimated on each superpixel, and the GGE light source map containing different light source color estimates is generated by the generalized gray world estimation method,

[0074] In this way, the corresponding light source map is generated from the input image;

[0075] The generalized gray world estimation method is as follows:

[0076] This method assumes that the average color of the scene under the illumination of a white light source is gray, then the RGB color observed at the pixel coordinate x of the image is f(x), as shown in the following formula (1),

[0077] f(x)=∫ Ω e(λ,x)s(λ,x)c(λ)dλ(1),

[0078] Among them, Ω represents the spectrum of visible light, λ represents the wavelength of light, e(λ,x) represents the spectrum of the light source, s(λ,x) is ...

Embodiment 2

[0111] In addition to the following:

[0112] The first step is to generate the corresponding light source map from the input image:

[0113] First, the input image is divided into similar regions, namely superpixels, and then the color of the light source is estimated on each superpixel, and the IIC light source map containing different light source color estimates is generated by the inverse intensity chromaticity estimation method.

[0114] In this way, the corresponding light source map is generated from the input image;

[0115] The inverse intensity chromaticity estimation method is as follows:

[0116] This method is based on the physical process of object surface reflection, and the observed image intensity is considered to be a mixture of diffuse reflection and specular reflection, then the color intensity on the image pixel coordinate x channel c∈{R,G,B} is p c (x), as shown in the following formula (3),

[0117] pc(x)=∫ Ω (e(λ,x)s(λ,x)+e(λ,x))c(λ)dλ (3),

[011...

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 a detection method of spliced ​​and falsified images, which relates to image analysis, and is a detection method of spliced ​​and falsified images based on light source maps and pyramid feature maps. Two convolutional neural networks with the same structure are used to extract spliced ​​falsified images respectively. The multi-stage features of the image and its corresponding light source map are combined with multi-scale information to fuse and upsample two sets of multi-stage features to obtain a pyramid feature map, and pass different layers of the pyramid feature map through the region generation network to obtain tampering candidate regions. The fixed-size feature map is generated by ROI Align, and the fixed-size feature map is classified, bounding box regression and mask prediction, and finally the bounding box and pixel-level positioning of the tampered area are obtained, and the detection of the stitched and tampered image is completed, which overcomes the current situation. The image tampering features extracted by existing technologies are single and incomplete, and it is easy to ignore tampering targets with small areas, and it is impossible to achieve end-to-end pixel-level positioning defects.

Description

technical field [0001] The technical solution of the present invention relates to image analysis, in particular to a method for detecting spliced ​​and falsified images. Background technique [0002] In recent years, image editing software such as Photoshop has been developed rapidly, allowing people to tamper with images according to their own wishes, even to the extent of confusing real ones. If these tampered pictures are used by people with ulterior motives in military politics, the court will testify , Scientific research and other serious occasions will distort the truth, mislead the public, and have an adverse impact on society. Therefore, it is of great significance to study the image forensics technology whether the digital image has been tampered with. [0003] Splicing is the most common means of image tampering. It is to splice a certain part of one image or multiple images into another image to achieve the purpose of falsifying facts. After some post-processing...

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): G06T7/00G06T7/11G06T7/90
CPCG06T7/0002G06T2207/20081G06T2207/20084G06T2207/20221G06T7/11G06T7/90
Inventor 于明刘世坤朱叶刘依郭迎春郝小可于洋师硕阎刚
Owner HEBEI UNIV OF TECH
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