Research On IPTV Video Content Analysis And Filtering

From the day of IPTV was born, it has indicated that IPTV would be a revolution of the information technology industry. IPTV contains the telephone, cable TV and broadband triple-play together, it can provide families with a variety of digital interactive network services. With the internet technology quick changes, IPTV network will provide customers and program provider with a wider emerging market. A key feature of IPTV is to provide users with high-quality and security information services, how to analyze the content situation of IPTV efficiently, fast and generally, and then ensure the program quality, becomes a very important investigation aspect, IPTV video content transfer is an important part of information transmission. Due to network bandwidth limitations and potential risks of network securities, how to do the IPTV video content analysis and filtering, becomes a key of information transmission.Traditional video content security technology is primarily conditional access and digital rights management technology. The former focuses on the area of digital TV conditional acceptance, and the latter focuses on copyright protection of digital content, both of them do not consider the regulation of video content, IPTV has special requirements on the video content analysis and filtering, it is difference with traditional video content analysis and filtering techniques.Based on the video content of IPTV and data transmission characteristics, this article specifies the related technologies of IPTV video information transmission, analyzes how to obtain the IPTV video data, how to deploy the network probe, how to manage the content analysis. On dealing with the video content, we are mainly focusing on the video data analysis and filtering, obtaining audio and parsing it. On the video data parsing aspect, we analyzed the illegal audio signal, character word definition, classification and matching and segmentation algorithm. Because of the support by existing Chinese word, speech recognition and classification algorithm, it is better for us to reduce the program audio level to be text dimension.The first thing considered of obtaining the IPTV data is the location, the IPTV system structure has following parts:the content network layer, the bearer network, home network. The bearer network and home network of this system will handle the application required by the customer to the system to be stream. During the process of data transmission, the data will be attacked; therefore, it is the position where to be considered to deploy the network probe, to obtain the data, the idea of obtaining the data is parsing and peeling off the data from the stream which is transferred in the network, PTV systems typically use a custom port to send streaming media data, transfer files with the ordinary way, this is the same, but the network application protocol header have their own characteristics, characteristics can be extracted as a basis for access to streaming media data to be analyzed, such as MPEG-TS packet every 188 bytes or 204 bytes will appear sync 0x47.In many cases, using one or more keywords with the full description of the video image of the wealthy information is very difficult, and this description there has great individual subjectivity. Customers are also difficult to use one or more key words to express their wishes clearly, and the customers' self-expression and the description of information are also difference. Therefore, the video image information transferred into text information is more difficult. But most of video data and sound is matched, so the audio information transferred into text and then the keyword matching the text can be achieved with traditional methods of audio and video search monitoring the content of public opinion.IPTV video content management can use the reverse positioning technology, the so-called reverse positioning is, after peel off the IPTV separate audio and video stream, the audio stream, if the current analysis of the IPTV video content as illegal, then the current audio stream through the analysis of the time stamp to be illegal, the customer can define standard IPTV video streaming, using standard video and audio streams replace illegal video stream.During the experiment, this validated the use of the audio signal analysis and public opinion analysis, which provides the transmission of video information a healthy, secure platform, and be able to take certain measures to manage illegal IPTV video and channel. In this thesis, according to the design of IPTV analysis and filtering process, only information for the IPTV signal analysis and filtering of audio is used, and did not consider the complex case of video content, therefore, it is the future investigation point of analyzing frequently changing video of IPTV. Through continued research, IPTV video analysis and filtering will be a major breakthrough in the coming days. The research work obtains Huawei Technologies Co., Ltd. (TS (Transport Stream) index prototype implementation) project support.