Evaluation of cost function for the oct tree segmentation of video cube

Segmentation is the pre-processing step of the various video processing applications like compression, foreground and background segmentation etc. It is a method of partitioning a video into uniform regions. Oct tree segmentation method which is explained in this paper assumes video as a cube and segments entire video cube into eight shorter size video cube. Segmentation has been formulated with the help of cost function. Whether the region is uniform or not is decided by the cost function. Cost function estimates uniform and non-uniform regions of video cube and if region is non-uniform, then they split otherwise they merge. A good selection of cost function is the aim of this paper. SPREF model has been used for the estimation of the cost function and it gives energy based cost function. The advantage of SPREF based cost function is that it incorporates both the texture properties and the motion of frames. performance of the proposed method has been evaluated with the other cost function and improved result obtained from proposed method validates its effectiveness.