LAUSR

With the tremendous increase in low resolution videos on video sharing websites, retrieval of a correct video becomes a tougher task. The existing methods provide retrieval approaches based on minimum number of features comparison. It leads to an inefficient video retrieval. Most researches had concentrated on tracking ability and conversion of low resolution to high resolution videos. These methods failed to provide fast retrieval of videos from large databases. The proposed work is concentrated mostly on riot videos from large video repositories to identify the previous criminal records in a particular region of the smart city (Cocchia in Smart and digital city: a systematic literature review, Springer International Publishing, Switzerland, 2014; Pardo and Taewoo in Proceedings of the 12th annual international conference on digital government research, ACM, New York, 2011). It uses certain combination ofobject oriented features like object and camera motion feature, color histogram and edge detection technique. In the proposed retrieval process, the key frames are extracted from the original video instead of using the whole video information for retrieval process. Object Oriented features were then extracted from these key frames and saved in database. Then, the retrieval process is done by searching the availability of relevant Object Oriented values based on the query submitted by the user. Thus the combination of four different features provides an efficient retrieval of low resolution videos from the database. The retrieved video may include redundant information in the projected work. To avoid such redundancy, particle swarm optimization (PSO) is used. The result of query video is compared with database video using degree of closeness measurement. Consequently, low resolution video retrieval based on PSO seems to be encouraging in terms of its performance in extracting videos than existing retrieval approaches.