Task–Adapted Learnable Embedded Quantization for Scalable Human-Machine Image Compression
IEEE Transactions on Circuits and Systems for Video Technology · 2025-05 · 35 (5) · 4768-4783
DOI: 10.1109/tcsvt.2025.3525664 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Circuits and Systems for Video Technology, 2025-05. Verfasst von Shaohui Li und 7 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 74 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Program of Shanghai Science and Technology Innovation Project, China Postdoctoral Innovation Talents Support Program.
Abstract (Originalsprache)
Image compression for both human and machine vision has become prevailing to accommodate to rising demands for machine-machine and human-machine communications. Scalable human-machine image compression is recently emerging as an efficient alternative to simultaneously achieve high accuracy for machine vision in the base layer and obtain high-fidelity reconstruction for human vision in the enhancement layer. However, existing methods achieve scalable coding with heuristic mechanisms, which cannot fully exploit the inter-layer correlations and evidently sacrifice rate-distortion performance. In this paper, we propose task-adapted learnable embedded quantization to address this problem in an analytically optimized fashion. We first reveal the relationship between the latent representations for machine and human vision and demonstrate that optimal representation for machine vision can be approximated with post-training optimization on the learned representation for human vision. On such basis, we propose task-adapted learnable embedded quantization that leverages learnable step predictor to adaptively determine the optimal quantization step for diverse machine vision tasks such that inter-layer correlations between representations for human and machine vision are sufficiently exploited using embedded quantization. Furthermore, we develop a human-machine scalable coding framework by incorporating the proposed embedded quantization into pre-trained learned image compression models. Experimental results demonstrate that the proposed framework achieves state-of-the-art performance on machine vision tasks like object detection, instance segmentation, and panoptic segmentation with negligible loss in rate-distortion performance for human vision.
Förderung
- National Natural Science Foundation of China
- Program of Shanghai Science and Technology Innovation Project
- China Postdoctoral Innovation Talents Support Program
Zitieren
APA
Li, S., Ma, S., Dai, W., Kan, N., Cheng, F., Li, C., Zou, J., & Xiong, H. (2025). Task–Adapted Learnable Embedded Quantization for Scalable Human-Machine Image Compression. IEEE Transactions on Circuits and Systems for Video Technology, 35(5), 4768-4783. https://doi.org/10.1109/tcsvt.2025.3525664
Vancouver
Li S, Ma S, Dai W, Kan N, Cheng F, Li C, et al. Task–Adapted Learnable Embedded Quantization for Scalable Human-Machine Image Compression. IEEE Transactions on Circuits and Systems for Video Technology. 2025;35(5):4768-4783. doi:10.1109/tcsvt.2025.3525664
BibTeX
@article{Li2025,
author = {Li, Shaohui and Ma, Shuoyu and Dai, Wenrui and Kan, Nuowen and Cheng, Fan and Li, Chenglin and Zou, Junni and Xiong, Hongkai},
title = {Task–Adapted Learnable Embedded Quantization for Scalable Human-Machine Image Compression},
journal = {IEEE Transactions on Circuits and Systems for Video Technology},
year = {2025},
volume = {35},
number = {5},
pages = {4768--4783},
doi = {10.1109/tcsvt.2025.3525664}
}
EUVIMED