Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Multiple pornographic image classification method based on image segmentation algorithm and deep learning

A deep learning and image segmentation technology, applied in image analysis, image enhancement, image data processing, etc. Diversity is difficult to establish and improve the accuracy of the skin color model, so as to improve the precision and recall rate, reduce the misjudgment and omission rate, and improve the achievability.

Inactive Publication Date: 2017-11-28
BEIJING ACT TECH DEV CO LTD
View PDF9 Cites 33 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0025] The first point is that the traditional pornographic image recognition algorithm calculates the proportion of skin color in the image to calculate the degree of nudity, which is limited by its theoretical basis, the diversity of the image environment and light, and the diversity of races. It is difficult to establish a perfect skin color model and improve the accuracy of detection;
[0026] The second point is that pornographic image recognition algorithms based on deep learning require a large number of image samples, which cannot adapt to the characteristics of diversification and differentiation of image content in network image formats. Greatly reduces the generalization ability and robustness of the system

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
  • Multiple pornographic image classification method based on image segmentation algorithm and deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0054] refer to figure 1 The improved image segmentation algorithm combined with deep learning multiple bad picture classification method includes four main steps: skin color recognition S1, principal component analysis of skin color area S2, deep learning S3, pornographic picture recognition based on convolutional neural network S4;

[0055] Step S1 skin color recognition:

[0056] (1) Convert the image to the YCbCr color space: Skin color detection is mainly based on the distribution characteristics of skin color in the color space to detect the skin color area in the image, because the YCbCr space can separate the brightness and chroma, and the CgCr chroma is affected by the brightness The impact of changes is less, and it is a two-dimensional independent distribution, so the skin color model is constructed in YCbCr chromaticity space;

[0057] (2) Use the expression (Cb > 77 And Cb 133 And Cr < 173) to traverse each pixel of the picture to detect whether the pixel color ...

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

A multiple pornographic image classification method based on an image segmentation algorithm and deep learning relates to the technical field of information, especially to the technical field of image identification. The method is characterized by comprising the following four major steps: skin color identification, principal component analysis of the skin color region, deep learning, and pornographic image identification based on a convolutional neural network. First, non-pornographic images are screened out through a skin color pixel detection and skin region partitioning algorithm based on the YCbCr theory. Undetermined images are input to a LeNet5-based convolutional neural network model for identification after feature extraction. Compared with the traditional identification based on skin color and features, the method can eliminate the noise influence of non-body-part images, is not constrained by light and human postures, and can greatly improve the accuracy of traditional pornographic image classification. Compared with general deep learning based on a convolutional neural network, the method does not need massive labeled images, and the characteristics of a deep residual network determine that the model can better analyze the characteristics of pornographic images. Only through about ten hours of training, an identification effect above 90% can be achieved.

Description

technical field [0001] The present invention relates to the field of information technology, especially the technical field of image processing. Background technique [0002] At present, most bad image recognition methods can be roughly divided into three categories according to their recognition objects: traditional skin color detection, feature detection and emerging pornographic image recognition based on neural networks. [0003] The traditional pornographic image recognition algorithm, which calculates the proportion of skin color in the image to calculate the degree of nudity, is limited by its theoretical basis, the diversity of the image environment and light, and the diversity of races make it difficult to establish a perfect Skin color model and improvement of detection accuracy, application number 200410042877.3 "a pornographic image detection method", which uses the ratio of the area of ​​the face area to the skin color area as the basis for judging pornographic ...

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): G06K9/62G06T7/90G06T3/40G06T7/11G06T7/13G06N3/04
CPCG06T3/40G06T7/11G06T7/13G06T7/90G06T2207/20081G06T2207/30196G06N3/045G06F18/2411
Inventor 林飞潘练赵喜荣熊骁毛俊
Owner BEIJING ACT TECH DEV CO LTD
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
Eureka Blog
Learn More
PatSnap group products