Super-High-Fidelity Image Compression via Hierarchical-ROI and Adaptive Quantization

arXiv:2403.13030v1 Announce Type: cross
Abstract: Learned Image Compression (LIC) has achieved dramatic progress regarding objective and subjective metrics. MSE-based models aim to improve objective metrics while generative models are leveraged to improve visual quality measured by subjective metrics. However, they all suffer from blurring or deformation at low bit rates, especially at below $0.2bpp$. Besides, deformation on human faces and text is unacceptable for visual quality assessment, and the problem becomes more prominent on small faces and text. To solve …

abstract aim arxiv compression cs.cv eess.iv fidelity generative generative models hierarchical however image low metrics progress quality quantization roi type via visual