Tagging and Clustering of Objects in Videos

“Tagging” in videos is an extension of the image tagging application. Just as people can be tagged in images, a system can be developed to carry out tagging in video clips. Using this, people can automatically be identified (“tagged”) wherever they are present in a video clip. The paper discusses the design of the system which identifies the real world objects in a video clip. The object to be searched can be provided by the user in the form of an image. The object is specified by a user outlining it in an image, and the object is then delineated in the retrieved shots. The aim is to detect and track presence of required objects in a video and for thesearching to be view-point invariant.Finally, the searched object can be tracked throughout the video by clustering all of its occurrences into a single video. This can be used for analyzing any possible changes the object might have undergone during the length of the video. General Terms Object detection, object matching, tagging, video analysis, image processing.