Object Classification for Video Surveillance System
Anjan Kumar Paul, Jae-Hong Min, Mohammad Khairul Islam, Young Bum Kim, Joong-Hwan Baek · ICEIC : International Conference on Electronics, Informations and Communications · 2010
Object classification has major applications in image retrieval, video summary generation, video surveillance etc. Video surveillance system possesses an important role in the field of security. Video surveillance is a major application where object classification having a vital role. The major drawback of current video surveillance system is that can not prevent intrusion without human interactions. Most of them are semi automated and decision making task mostly depend upon human interactions. Existing systems also deal with the single class, mostly with human class. Classification and recognition of left behind objects is very important for smart video surveillance system. We have developed a smart object classification module for current video surveillance system where we adopted the concept of bag of words. The total work we have finished in two major steps, training which is accomplished in offline, testing is finished at online with still image. We test with four different classes of dataset dividing them into test and training samples and our overall performance is good.