Video Encoding Apparatus Resolution Varying Motion Estimation

The encoding process is segmented into different resolution levels. The patent divides the video sequence into multiple frames and applies different resolution levels to different frames based on their importance. Important frames (e.g., I-frames, key P-frames) are encoded at full resolution, while less important frames are encoded at reduced resolution, thereby reducing overall computational complexity while maintaining encoding quality for critical content.

Different quality levels are applied locally to different frames within the video sequence. The patent uses a resolution control signal to selectively apply high resolution to specific frames that require it (such as frames with significant motion or important visual information) and low resolution to other frames, optimizing the balance between quality and computational load on a frame-by-frame basis.

The patent segments the video encoding process by applying motion estimation and compensation selectively to different frames rather than uniformly to all frames. Frames are divided into groups where only certain frames undergo full motion estimation and compensation, while other frames use simplified or skipped motion compensation, thereby reducing total processing time while maintaining encoding accuracy for frames that require it.

The patent applies partial motion estimation and compensation action to only the necessary frames rather than performing full motion estimation on every frame. By identifying and processing only the frames that require accurate motion compensation (based on frame type, motion activity, or other criteria), the system achieves sufficient encoding accuracy while significantly reducing processing time and avoiding unnecessary computational delays.