Progressive Learned Image Compression for Machine Perception

TL;DR

This paper introduces PICM-Net, a progressive learned image compression method tailored for machine perception, featuring trit-plane coding and an adaptive decoding controller to optimize transmission and classification accuracy.

Contribution

The work presents a novel progressive image compression codec for machines, incorporating trit-plane coding and an adaptive decoder for improved efficiency and performance.

Findings

Enables efficient progressive transmission for machine perception

Maintains high classification accuracy with adaptive decoding

Establishes a new paradigm for machine-aware image compression

Abstract

Recent advances in learned image codecs have been extended from human perception toward machine perception. However, progressive image compression with fine granular scalability (FGS)-which enables decoding a single bitstream at multiple quality levels-remains unexplored for machine-oriented codecs. In this work, we propose a novel progressive learned image compression codec for machine perception, PICM-Net, based on trit-plane coding. By analyzing the difference between human- and machine-oriented rate-distortion priorities, we systematically examine the latent prioritization strategies in terms of machine-oriented codecs. To further enhance real-world adaptability, we design an adaptive decoding controller, which dynamically determines the necessary decoding level during inference time to maintain the desired confidence of downstream machine prediction. Extensive experiments demonstrate that our approach enables efficient and adaptive progressive transmission while maintaining high performance in the downstream classification task, establishing a new paradigm for machine-aware progressive image compression.