Comprehensive Comparisons of Uniform Quantization in Deep Image Compression
Open Access
- 11 January 2023
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 4455-4465
Abstract
In deep image compression, uniform quantization is applied to latent representations obtained by using an auto-encoder architecture for reducing bits and entropy coding. Quantization is a problem encountered in the end-to-end training of deep image compression. Quantization’s gradient is zero, and it cannot backpropagate meaningful gradients. Many methods have been proposed to address the approximations of quantization to obtain gradients. However, there have not been equitable comparisons among them. In this study, we comprehensively compare the existing approximations of uniform quantization. Furthermore, we evaluate possible combinations of quantizers for the decoder and the entropy model, as the approximated quantizers can be different for them. We conduct experiments using three network architectures on two test datasets. The experimental results reveal that the best approximated quantization differs by the network architectures, and the best approximations of the three are different from the original ones used for the architectures. We also show that the combination of quantizers that uses universal quantization for the entropy model and differentiable soft quantization for the decoder is a comparatively good choice for different architectures and datasets.Keywords
Funding Information
- Hitachi, Ltd
- Japan Society for the Promotion of Science (JSPS) Kakenhi (22J13735)
This publication has 31 references indexed in Scilit:
- Image Super-Resolution Using Very Deep Residual Channel Attention Networks2018
- Deep Image Compression with Iterative Non-Uniform Quantization2018
- Non-local Neural Networks2018
- Full Resolution Image Compression with Recurrent Neural Networks2017
- Deep Residual Learning for Image Recognition2016
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Efficient scalar quantization of exponential and Laplacian random variablesIEEE Transactions on Information Theory, 1996
- On universal quantization by randomized uniform/lattice quantizersIEEE Transactions on Information Theory, 1992
- The JPEG still picture compression standardCommunications of the ACM, 1991
- On universal quantizationIEEE Transactions on Information Theory, 1985