TACTIC: Joint Rate-Distortion-Accuracy Optimisation for Low Bitrate Compression
This addresses the challenge of efficient image compression for downstream AI tasks, offering incremental improvements over existing compression techniques.
The paper tackles the problem of optimizing lossy compression for specific tasks by jointly considering rate-distortion-accuracy trade-offs, resulting in a 4.5% accuracy improvement on ImageNet classification and 3.4% accuracy and 4.9% mean IoU gains for semantic segmentation compared to task-agnostic methods at low bitrates.
We present TACTIC: Task-Aware Compression Through Intelligent Coding. Our lossy compression model learns based on the rate-distortion-accuracy trade-off for a specific task. By considering what information is important for the follow-on problem, the system trades off visual fidelity for good task performance at a low bitrate. When compared against JPEG at the same bitrate, our approach is able to improve the accuracy of ImageNet subset classification by 4.5%. We also demonstrate the applicability of our approach to other problems, providing a 3.4% accuracy and 4.9% mean IoU improvements in performance over task-agnostic compression for semantic segmentation.