# On Disentangled Training for Nonlinear Transform in Learned Image Compression

arXiv: https://arxiv.org/abs/2501.13751 · HTML: https://mlsanity.com/paper/2501.13751/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Jan 2025.

## Built on by (most cited first, 9 of 9)

- [Content-Aware Mamba for Learned Image Compression](https://mlsanity.com/paper/2508.02192/) (2025)
- [Leveraging Second-Order Curvature for Efficient Learned Image Compression: Theory and Empirical Evidence](https://mlsanity.com/paper/2601.20769/) (2026)
- [High Quality Underwater Image Compression with Adaptive Color Correction](https://mlsanity.com/paper/2505.09986/) (2025)
- [Uni-LVC: A Unified Method for Intra- and Inter-Mode Learned Video Compression](https://mlsanity.com/paper/2603.05756/) (2026)
- [Adaptive Learned Image Compression with Graph Neural Networks](https://mlsanity.com/paper/2603.25316/) (2026)
- [Error-Propagation-Free Learned Video Compression With Dual-Domain Progressive Temporal Alignment](https://mlsanity.com/paper/2512.10450/) (2025)
- [MedROI: Codec-Agnostic Region of Interest-Centric Compression for Medical Images](https://mlsanity.com/paper/2604.04511/) (2026)
- [ChWDTA: Channel-wise Wavelet-Domain Transformer Attention and Entropy Modeling for Learned Image Compression](https://mlsanity.com/paper/2606.00111/) (2026)
- [Feed-Forward Multi-View Feature Coding for 3D Gaussian Splatting](https://mlsanity.com/paper/2610.20590/) (2026)

## Abstract (arXiv)

Learned image compression (LIC) has demonstrated superior rate-distortion (R-D) performance compared to traditional codecs, but is challenged by training inefficiency that could incur more than two weeks to train a state-of-the-art model from scratch. Existing LIC methods overlook the slow convergence caused by compacting energy in learning nonlinear transforms. In this paper, we first reveal that such energy compaction consists of two components, i.e., feature decorrelation and uneven energy modulation. On such basis, we propose a linear auxiliary transform (AuxT) to disentangle energy compaction in training nonlinear transforms. The proposed AuxT obtains coarse approximation to achieve efficient energy compaction such that distribution fitting with the nonlinear transforms can be simplified to fine details. We then develop wavelet-based linear shortcuts (WLSs) for AuxT that leverages wavelet-based downsampling and orthogonal linear projection for feature decorrelation and subband-aware scaling for