Research On Technologies Of Capturing Image In Network Erotic Image Filtering

Popularization of Internet brings enormous convenience and the ample information to people's lives. Meanwhile it is also been seen that massive pornographic, reactionary, unwholesome information is disseminated recklessly by virtue of Internet's openness, sharing and the connectivity. At present the youth are the majority of the netizens in our country, and their age structure displays a characteristic of becoming gradually younger. For the sake of weak discrimination and resistivity of the youths, the network pornography addicts some youths to it, induce them to abandon study and to become trance, decadence. Accordingly, plenty of paterfamilias and pedagogues are anxious for the youths. And we cannot ignore the poison which leads adults to crime of pornographic website either. Obscene pornography's poisoning is much like the narcotics which would make huge injury especially to the youths who have weak control ability and have not formed their world outlook and the values.The pornography which is harmful for the youth has already become a social effect of pollution. How to prevent the young people from accessing the pornography content is a serious social problem. And it is a problem that has the most attention of the parents. At present most of the countries have formulated laws against the pornography over internet, and make necessary actions for filtering.At present there are several methods such as text-based filter, URL-based filter and image-based filter for filtering the unwholesome information. More and more scholars and organizations pay attention to the research of image-based filter because the images have more information and stronger feeling view stimulation than the text. The work of Filtering network adult image is analyzing and detecting image, and the precondition is the capture of image data. Founded on the project"Research on Content-Based Erotic Image Filtering technique and its realization in IE"of Zhuhai Science and Technology Planning Projects in 2004, in this paper, it studies the key technology of network image capture and builds two filters with different principle. The main work of the dissertation is as follows:(1) The research of image capturing technology in Windows platform. This paper introduces the network characteristic of Windows platform, and analyzes key technique of network data filtering in Windows platform emphasis on the application layer and kernel layer. After that, this paper discusses Internet Explorer plug-in extension technology and Intermediate driver filtering in Network Driver Interface Specification.(2)The Erotic image filtering of Internet Explorer based on its plug-in extension technology.The Internet Explorer architecture's foundation is the Composition Object Model. Microsoft has provided very good technical support for the extension of IE function. BHO (Browse Helper Object) and MIME Filter (Multipurpose Internet Mail Extension protocol Filter) are some of extension technology for IE. This paper analyzes the principium of BHO and MIME Filter, and designs a scheme about IE adult image filter based on BHO&MIME Filter. The filter mainly includes four parts: the BHO module, the MIME Filter module, the image analyzing module and the function setup module. The first module, the BHO plug-in unit, takes charge of the filter's loading, initialization and the Erotic image deletion. The MIME Filter module takes charge of the blacklist modifying and the image data capturing and packaging. The detecting module and the setup module analyze image and complete the filter's setup. The scheme solves some questions which appear in algorithm such as blinking, refreshing and so on. This filter is integrated into IE seamlessly, and it is good at real-time and stability.(3)The design and simulation of image filter based on IMDThe IMD works in the kernel layer. It is one kind of technology in the network security product. This paper introduces the characteristics of kernel driver and the NDIS, analyzes the process of receiving and dispatching of IMD data, and proposes scheme of image filter based on IMD. The scheme includes two parts: the kernel module and the application module. The kernel module, the IMD driver, captures packets and builds a linked list for the image requested, recombines image packages according to the image reorganization algorithm, notifies application layer module receiving and detecting data by event mechanism. And it has delay timer for notifying the upper level to manage the processed image data. The application module takes charge of the communicating and the sharing of stored data between application layer and IMD. It detects the packaged image data and modifies the data flag bit according to the analysis processing result. In view of the difficult of the development and debugging of IMD, this paper proposes a simulant algorithm for IMD. This paper analyzes selectively the related content about TCP/IP and HTTP, and presents the recombining algorithm of image data packets. According to the Hash algorithm used in Linux's NetFilter and using the insertion sort based on the network data characteristic, this algorithm has a good operating speed, and has satisfied the request of real-time capture.There are still many aspects of our two filtering systems that need to be improved and perfected. The filter of IE wastes much system resource and can be used only on browsers that based on IE core. The filter of IMD only completes the major algorithm in user-mode and needs explanting to kernel. These are also our future work.