International Journal of Engineering Research and Applications (IJERA) ISSN: 2248-9622 International Conference on Industrial Automation And Computing (ICIAC- 12 th & 13 th 2014) Jhulelal Institute of Technology, Nagpur 25 | Page Rakhi Ashok Aswani*, Shailesh D.Kamble** *(Department of Computer Technology, Yeshwantrao Chavan College of Engineering, Nagpur Email: rakhi7.aswani@gmail.com) ** (Department of Computer Technology, Yeshwantrao Chavan College of Engineering, Nagpur Email: shailesh_2kin@rediffmail.com) ABSTRACT The large usage of multimedia applications on internet and mobiles has increased the demand of compressed data, in order to reduce the requirement of bandwidth and time to transfer the data. So block based method for video compression is getting more importance due to its effectiveness and easy implementation. Fractal video compression follows the property of self-similarity. That's the biggest reason for high compression ratio and also attracts more researchers to work for fractal compression. In block matching motion estimation various algorithm has been proposed having different search pattern and strategies. As the search pattern gives large impact on efficiency of the algorithm. The paper presents an efficient method on fractal video compression, which follows half-way stop technique. Keywords - bin-tree partitioning, fractal video compression, inter-pixel similarity, intra-pixel similarity , quad- tree partitioning, bin-tree partitioning I. INTRODUCTION Block matching motion estimation becomes the most attractive area for research. If the frame rate is high, then there will be great amount of similarity in successive frames. The temporal redundancy needs to be eliminated between successive frames. The frames are divided into rectangular blocks and block displacement need to be calculated from previous frame with the current frame. This process is computational intensive and takes 80% of the encoding time. The concept of block matching is based on matching of a block with other blocks. So the image is partitioned into domain blocks and range blocks. Domain blocks are non-overlapping blocks, while range blocks are overlapping blocks. As an image can have some similar blocks also. It is not necessary that all the blocks in an image are completely dissimilar with each other. So main task is finding those blocks in entire video by using intra-pixel and inter-pixel property, to save the storing space as well as for compression purpose. Intra-pixel similarity refers to similarity in successive frames of a video while in inter-pixel similarity refers to similarity within a frame only. An image can be partitioned into blocks by bin-tree partitioning or Quad-tree partitioning or triangular partitioning, Where bin-tree partitioning some times refers to horizontal vertical partitioning. While in Quad-tree partitioning the image is divided into four quadrants. the partitioning is continued till no dissimilarity remained in image. We have proposed an efficient Four-Step Search, which gives better results than FS, 3SS, 4SS & DS. The rest of the paper is organized as follows. The second section explains about fractal video compression. The third section explains 4SS. The fourth section explains E4SS. Then experimental results are shown. II. FRACTAL VIDEO COMPRESSION The fractals uses the concept of self similarity. Fractal image compression reduces the redundancy of images by using self-similarity property and is a favorable method for compressing an image due to its high compression ratio, fast decompression and resolution independence[4]. The process of Fractal Video Compression is same as Fractal Image Compression. The major difference between them is that, video compression takes advantage of the similarity between successive frames and hence we obtain higher compression rates, if compared with image compression[5]. An Efficient Method for Fractal Video Compression using Block Matching Motion Estimation RESEARCH ARTICLE OPEN ACCESS