Velocity and Direction-Based Motion Estimation in Video Compression to Reduce Computation and Increase Video Quality

<p style="text-align: left; padding-right: 30px;">High-Efficiency Video Coding (HEVC) represents the latest advancement in video compression standards, offering significantly improved efficiency by supporting high-definition (HD) videos with double the compression capability of the H.264 standard. However, motion estimation within HEVC remains the most computationally intensive aspect of the encoding process, accounting for the majority of the required processing time. Although various methods have been developed to mitigate this complexity, and many have demonstrated practical effectiveness, the processing demands continue to pose challenges for real-time applications. This research introduces a fast sub-pixel motion estimation algorithm designed to minimize the number of search points during video encoding. The proposed approach leverages a mathematical model grounded in motion physics and image characteristics across consecutive frames, incorporating statistical insights into the movement of elements within the video sequence. By reducing the number of search points, the algorithm lowers computational complexity while maintaining or even enhancing relevant video quality metrics.