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

Discrimination method between natural image and computer-generated image based on statistical properties of dct domain

A technology for natural images and image generation, which is applied in the field of image identification and can solve the problem of low identification accuracy.

Inactive Publication Date: 2016-02-17
BAINIAN JINHAI SCI & TECH +1
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] Aiming at the problem that the identification accuracy of the current natural image and computer-generated image identification method is not high, a method for identification of natural image and computer-generated image based on the statistical characteristics of DCT domain is proposed.

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
  • Discrimination method between natural image and computer-generated image based on statistical properties of dct domain
  • Discrimination method between natural image and computer-generated image based on statistical properties of dct domain
  • Discrimination method between natural image and computer-generated image based on statistical properties of dct domain

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0032] The basic principle of this method is to first perform Gaussian blur processing on the image, and then use Benford's model to analyze the probability distribution characteristics of the first significant digit of the DCT domain AC coefficient of the three RGB color channels of the image, and finally classify and identify the image according to the different image characteristics . The following is a detailed introduction to the benford model, Gaussian blur, and DCT domain statistical characteristics in the algorithm:

[0033] 1Benford model

[0034] Benford's law is also known as "First-digitphenomenon", Significantdigitlaw, and LogarithmLaw. It is to detect the inner law of the little-known number distribution from the statistical point of view. The law reveals that when certain conditions are met. In a large number of statistical data, the probability distribution law of the number 1-9 appearing at the first place in the data.

[0035] In the 1930s, Frank Benford,...

Embodiment 2

[0050] Algorithm flow chart of the present invention is as figure 1 As shown, the implementation steps of the identification method are as follows:

[0051] (1) Firstly, Gaussian blur processing and dimension reduction processing are performed on the image to be tested, and then 8×8 non-repetitive block DCT discrete cosine transform is performed on the three channels of the image R, G, and B, respectively, and the obtained 8×8 block DCT coefficient matrix .

[0052] In the Gaussian blur processing, the image is subjected to Gaussian blur processing, the blur radius is set to 0.3, and the normal distribution equation in two-dimensional space is:

[0053] G ( r , σ ) = 1 2 π σ 2 e - r 2 ...

Embodiment 3

[0065] The technical solution adopted by the present invention to solve the above-mentioned technical problems is: a method for discriminating between natural images and computer-generated images based on the statistical properties of the DCT domain, and the method for determining the threshold T in the discriminating method is as follows:

[0066] (1) Firstly, Gaussian blur processing and dimensionality reduction processing are performed on the images of the experimental group, and then 8×8 non-repetitive block discrete cosine transforms are performed on the RGB3 channels of the image to obtain an 8×8 block DCT coefficient matrix.

[0067] (2) Make statistics on the distribution of the first significant digit of the DCT domain AC coefficient of each channel, and obtain three probability distribution curves.

[0068] (3) Calculate the average absolute difference of the three curves, compare the maximum and minimum values ​​of the average absolute difference by counting the aver...

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 discloses a method for identifying a natural image and a computer generated image based on DCT (Discrete Cosine Transformation)-domain statistic characteristics. The method is characterized by comprising the following steps of (1) firstly, carrying out gaussian fuzzy processing and dimension reduction processing on an image to be detected, carrying out 8*8 no-repeat DCT on3 channels R, G and B of the image to obtain a 8*8 partitioning DCT coefficient matrix; (2) carrying out statistics on the distribution of a first significant figure of the DCT-domain AC coefficient of each channel to obtain three probability distribution curves; and (3) calculating an average absolute differential of the three probability distribution curves, if the average absolute differential is greater than a set threshold T to prove that the overlapping degree of the three curves is not sufficient, judging the image to be detected as the natural image, otherwise, judging the image to be detected as the computer generated image. Experimental results show that due to the algorithm, the identification accuracy rate of the natural image and the computer generated image is increased. Compared with the existing algorithm, the method has the advantages that the identification rate is higher, the calculating amount is small, the implementation is easy, and the identification accuracy rate reaches 95.22%.

Description

technical field [0001] The invention relates to the technical field of image identification, in particular to a method for identifying natural images and computer-generated images based on DCT domain statistical properties. Background technique [0002] With the advent of the information age, digital imaging equipment is becoming more and more common, and traditional film images are gradually being replaced by digital images, but digital images are easier to be tampered with and forged. Some 3D image generation software can easily generate computer images that are so fake that the human eye can hardly distinguish them from natural images. With the continuous expansion of the application range of digital images, especially in news, judicial and other industries, how to identify their corresponding sources in the face of a large number of digital images has become an urgent problem to be solved. [0003] The identification method of real images and computer-generated images h...

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): G06T7/00
Inventor 陈长宝张震杨宇豪杜红民谢永杰佟森峰庄东刚盛铎宋超范秉琪赵晓祥崔帅
Owner BAINIAN JINHAI SCI & 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