A Dynamic Parallel and Pipelined Architecture for Intra Prediction in H.265 Standard

Abstract

In the present world where technology is growing faster, the video based applications are rapidly increasing and needs a technology which supports high resolution videos. High Efficiency Video Coding (HEVC) method is one which works on 4K and 8K video applications. In this work we have implemented the new parallel and a hardware accelerator which is highly efficient for the intra prediction blocks. Due to parallel and pipelined architecture, Intra Prediction speeds up the process of prediction and also minimizes the time required for accessing the data from the memory. The given architecture design reduces Area, Power and Delay elements. The results when compared with different FPGA versions shows that our architecture consumes 69 LUTs in ZYNQ FPGA for 4X4 pixels.