A NEW APPROACH FOR VIDEO OBJECT MINING: ISSUES AND CHALLENGES

In these days, due to the increasing rate of video data over the World Wide Web, it is becoming very essential to extract useful information from visual data. The video data consists of various objects as its important visual information. Also video data contains huge amount of direct as well as indirect/hidden information about objects. The users can access direct information from video by viewing it. To access hidden information from the video, mining techniques such as classifications, clustering, regression, outlier detection and association rules etc, can be applied to discover knowledge from the video data, however, there is no ideal framework to mine video objects effectively. In this paper we reviewed some existing video object mining techniques and proposed a new approach for effective and efficient video object mining framework to extract useful knowledge from videos. An attempt is also made to mention about some research challenges in video object mining.