Intelligent Alerting System Development by Network Media Semantic Analysis and Face Recognition

Abstract

Nowadays, the network has become the major data provider to everyone. Various providers feed data to the end users to achieve the flow of information for their purposes. In this research, we develop an intelligent alerting system that can automatically sematic analyze the contents from the network and extract the altered information including event description and suspect’s photo. We crawl the PTT, Dcard, and Google News in Taiwan as our sources to process due to the popularity of the three social network systems. These are the three very popular social media in Taiwan and real-time events, news is always posted to the three media very soon, therefore, we adopt them as our research source. Video monitors installed on campus or in the community after receiving alerts and photos will enter alert mode. Real-time face recognition will be issued for the people captured by the camera. Once the system recognizes the suspect, it will notify the security guard to take necessary actions to protect the people. There are two major parts to this research, the first one is the media semantic analysis and the second one is the real-time face recognition. It is an intelligent altering system that is capable to find out the suspected assault or rapist rapidly once he is captured by a monitor and recognized after people post the crime events to the media.