Objectionable image distinguishing method integrating skin color, face and sensitive position detection

Abstract

The invention relates to an objectionable image distinguishing method integrating skin color, face and sensitive position detection. The method comprises the following steps that a skin color model is firstly built, the face detection is carried out, the constituted feature vector of skin color and face features is extracted, a SVM (support vector machine) algorithm is utilized for training, and a SVM classifier is obtained; then, by aiming at the female breast in the local key position of the human body, SIFT (scale-invariant feature transform) features are extracted, an Adaboost algorithm is utilized for training, and an Adaboost classifier is obtained; next, by aiming at the female private parts in the local key position of the human body, the trunk region of the human body is determined, haar-like features are utilized as a template for carrying out searching and matching in the trunk region of the human body; and finally, the SVM classifier, the Adaboost classifier and the template matching method are adopted for carrying out image detection, a C4.5 decision-making tree method is utilized for integrating detection results, a decision-making tree model is built, the decision-making tree model is adopted for recognizing objectionable images, and the final distinguishing results are given. The objectionable image distinguishing method has the advantages that the detection accuracy is improved, and meanwhile, the execution speed is ensured.

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a bad image discrimination method which integrates skin color, human face and sensitive part detection. Background technique [0002] The development of science and technology in today's world is changing with each passing day, and the rapid development of Internet technology has greatly affected the daily life of human beings. The current Internet has covered every corner of the world, and various images, texts, and audio information have also spread to every corner of the world through the Internet. Some of this information has provided great help to human learning, work and life, but some information has also brought adverse effects, such as pornographic information. As the influence of the Internet continues to expand, many criminals spread pornographic information through Internet technology. In recent years, pornographic websites and pornographic webp...

AI Technical Summary

Problems solved by technology

However, there is a lack of detection of key parts of the human body. For some images that do not expose much skin color but expose key parts, it is easy to miss detection.

In the article "Study on Recognition of Key Parts of Sensitive Images" by Wang Shen of Lanzhou University, although a detection method for female breasts was proposed, based on the simple features of color and shape, the typical features of key parts of the human body were not discovered, and the false positive rate of recognition was high.

Zhou Jianzheng of Hangzhou Dianzi University "A Method for Filtering Bad Network Images Based on SVM" introduced the Support Vector Machine (SVM) algorithm in machine learning to extract features such as skin color, face, and shape. This method extracts higher-level images. feature, and adopts the method of machine learning, which improves the detection rate and reduces the false alarm rate, but does not propose the detection of key parts of the human body, which does not have a large area of ​​skin color in bad images, but pornographic images that expose key parts are easy to produce False report

In the article "Design and Implementation of Bad Image Detection System" by Hou Chun of Nanjing University of Science and Technology, a method of using haar features to detect female breasts and private parts of female lower body is proposed, but this method needs to search for key areas and filter windows in the entire image. Matching, not only slower but also less accurate

Images