Post shot boundary detection technique: flashlight scene determination

Shot boundary detection, or scene change detection, is a technique used in the initial phase of video indexing. One of the problems in detection is the discrimination of abrupt scene change from flashlight scenes. Current method of detecting flashlight is based upon object edges, in which performance can be affected by the contents of the scene. To overcome this, we present a novel method that utilises the edge direction, reducing erroneous matching with increasing dilation radius. This will improve the accuracy of similarity testing and reduce the amount of erroneously matched edges by at least four times. Our experiment in discriminating flashlight effects from abrupt scene change frame pairs shows that our technique produces a perfect detection, which cannot be achieved by feature-based detection. Such contribution is important as it improves the indexing of real life video used in the upcoming video database standard MPEG-7