Tackling Ill-posedness of Reversible Image Conversion with Well-posed Invertible Network
Lists
Only you see this
Abstract
Reversible image conversion (RIC) suffers from ill-posedness issues due to its forward conversion process being considered an underdetermined system. Despite employing invertible neural networks (INN), existing RIC methods intrinsically remain ill-posed as inevitably introducing uncertainty by incorporating randomly sampled variables. To tackle the ill-posedness dilemma, we focus on developing a reliable approximate left inverse for the underdetermined system by constructing an overdetermined system with a non-zero Gram determinant, thus ensuring a well-posed solution. Based on this principle, we propose a well-posed invertible $1\times1$ convolution (WIC), which eliminates the reliance on random variable sampling and enables the development of well-posed invertible networks. Furthermore, we design two innovative networks, WIN-Naïve and WIN, with the latter incorporating advanced skip-connections to enhance long-term memory. Our methods are evaluated across diverse RIC tasks, including reversible image hiding, image rescaling, and image decolorization, consistently achieving state-of-the-art performance. Extensive experiments validate the effectiveness of our approach, demonstrating its ability to overcome the bottlenecks of existing RIC solutions and setting a new benchmark in the field. Codes are available in https://github.com/BNU-ERC-ITEA/WIN.
BibTeX
@misc{huang2025tackling,
author = {Yuanfei Huang and Hua Huang},
title = {{Tackling Ill-posedness of Reversible Image Conversion with Well-posed Invertible Network}},
year = {2025},
eprint = {2508.02111},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2508.02111},
}| IICNet: A Generic Framework for Reversible Image Conversion | arXiv · 2021 |
| Invertible Image Rescaling | arXiv · 2020 |
| Invertible Rescaling Network and Its Extensions | arXiv · 2022 |
| Invertible Residual Rescaling Models | arXiv · 2024 |
| Large-Capacity Image Steganography Based on Invertible Neural Networks | CVPR · 2021 |
| Image processing network-inverted identifiable secure natural steganography | Journal of Image and Graphics · 2023 |