See Without Decoding: Motion-Vector-Based Tracking in Compressed Video
This enables real-time analytics for large monitoring systems by reducing computational overhead.
The paper tackled the problem of object tracking in compressed video without full decoding, achieving a computational speed-up of up to 3.7 times with only a 4% drop in mAP@0.5 compared to RGB baselines on MOTS15/17/20 datasets.
We propose a lightweight compressed-domain tracking model that operates directly on video streams, without requiring full RGB video decoding. Using motion vectors and transform coefficients from compressed data, our deep model propagates object bounding boxes across frames, achieving a computational speed-up of order up to 3.7 with only a slight 4% mAP@0.5 drop vs RGB baseline on MOTS15/17/20 datasets. These results highlight codec-domain motion modeling efficiency for real-time analytics in large monitoring systems.