Block-optimized Variable Bit Rate Neural Image Compression
This work addresses image compression efficiency for applications requiring variable bit rates, though it appears incremental in nature.
The authors tackled neural image compression by proposing a block-based auto-encoder system with multiple novel contributions, achieving incremental performance improvements as evaluated in their study.
In this work, we propose an end-to-end block-based auto-encoder system for image compression. We introduce novel contributions to neural-network based image compression, mainly in achieving binarization simulation, variable bit rates with multiple networks, entropy-friendly representations, inference-stage code optimization and performance-improving normalization layers in the auto-encoder. We evaluate and show the incremental performance increase of each of our contributions.