Abstract

Object segmentation is the process of abstraction of object from an image or video frame. There are several techniques available to handle object segmentation from video frames. Existing methods have mainly two problems such as they do not consider the spatial parameters of an object in a frame, and it is not possible to handle complex scenes with highly dynamic background movements with existing methods. With varying scales, a robust analysis mechanism is required to handle background regions or foreground movements. This paper proposes a solution to these problems and provide an efficient algorithm to overcome these problems. Also suggested method for object detection, identification and license plate identification from dynamic background videos.