Study of Video Compression Techniques for Low Bit Rate Applications

Abstract

In recent times, the video compression and communication play a vital role in many applications due to the demand of video, video-conferencing, remote education etc. The multimedia industry has enabled widespread usage of digital television, internet based applications and smart phones therefore, significance of video compression techniques has become prominent. Video compression is used to minimize the required number of bits to represent the data which leads to the savings in storage capacities, reduction in the hardware cost and availability of network bandwidth requirements. The process of video compression utilizes many coding methods to reduce the redundancy in video data. This research is aimed at comprehensive study of various techniques involved in video compression and understanding the possibility for lowering bit rate without compromising the quality of video. In this work, the most widely utilized video compression methods and international coding standards are analysed effectively for developing an efficient algorithm which significantly reduces the bit rate after compression without compromising the quality of video. An emphasis is given to video compression applications that are relevant for battery-powered, resource-constrained devices. To achieve the intended goal, Accordion algorithm along with Modified Run Length Encoding (RLE) technique and efficient handling of DC coefficients for creating Huffman dictionary in improvised way is proposed for inter and intra-frame compression. This method is used to optimize the quality of video performance to obtain optimum compression ratio and also to adapt the narrow bandwidth environment. This methodology provides better results compared to Audio Video Interleaving (AVI) standard. The proposed approach improves the features for better compression quality and reduces bit rate significantly. Further, an Octonary Repetition Tree (ORT) method is used to address duplication problem of RLE and to minimize the computational complexity. Physical-Next Generation