Content Based Video Retrieval using Mobile Agent

With the evolution of technology, large amount of information is available in heterogeneous formats and need to be retrieved efficiently in distributed environment. Content based multimedia retrieval system (CBMR) is a common approach based on automatic retrieval process, rather than traditional keyword-based technique, which was very time-consuming. The growing usage of network interrupts the working of network models and puts tremendous pressure on bandwidth requirements and network management. Any form of distribution brings with it the management overhead and the burden of dealing with it lies with the user. Finding and managing information over the network is one of the complex and time-consuming process. Mobile agent is an emerging technology of distributed computing that can overcome the barriers posed by network congestion and helps in getting response to a query. It acts on behalf of a user and fits well in the low-bandwidth network environments. Mobile agents are used for managing the network traffic, improving reliability, providing fault tolerance and also for multimedia retrieval. Research is being carried on the use of mobile agents for image retrieval. In this paper, mobile agent based architecture for content based video retrieval system is implemented. It focuses on searching and retrieving video files and maintaining video metadata database. The video comparison is based on calculation of feature vector. The paper shows the results of comparison of two videos by calculating the feature vector for every frame and comparing them frame by frame. After comparison the similar videos are retrieved from the remote machine and results of comparison are stored in the database on the source machine.