Post-Training Quantization for Cross-Platform Learned Image Compression
This solves a critical issue for deploying learned image compression in industrial applications by ensuring cross-platform consistency, though it is incremental as it builds on existing state-of-the-art models.
The paper tackles the problem of non-deterministic calculations in learned image compression, which cause cross-platform inconsistencies and decoding failures, by introducing post-training quantization to enable integer-only inference while maintaining superior rate-distortion performance.
It has been witnessed that learned image compression has outperformed conventional image coding techniques and tends to be practical in industrial applications. One of the most critical issues that need to be considered is the non-deterministic calculation, which makes the probability prediction cross-platform inconsistent and frustrates successful decoding. We propose to solve this problem by introducing well-developed post-training quantization and making the model inference integer-arithmetic-only, which is much simpler than presently existing training and fine-tuning based approaches yet still keeps the superior rate-distortion performance of learned image compression. Based on that, we further improve the discretization of the entropy parameters and extend the deterministic inference to fit Gaussian mixture models. With our proposed methods, the current state-of-the-art image compression models can infer in a cross-platform consistent manner, which makes the further development and practice of learned image compression more promising.