# Opening the Black Box of Learned Image Coders

arXiv: https://arxiv.org/abs/2202.13209 · HTML: https://mlsanity.com/paper/2202.13209/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 26 Feb 2022.

## Built on by (most cited first, 5 of 5)

- [Computationally-Efficient Neural Image Compression with Shallow Decoders](https://mlsanity.com/paper/2304.06244/) (2023)
- [Color Learning for Image Compression](https://mlsanity.com/paper/2306.17460/) (2023)
- [A Taxonomy of Miscompressions: Preparing Image Forensics for Neural Compression](https://mlsanity.com/paper/2409.05490/) (2024)
- [Few-Shot Domain Adaptation for Learned Image Compression](https://mlsanity.com/paper/2409.11111/) (2024)
- [Trustworthy Compression? Impact of AI-based Codecs on Biometrics for Law Enforcement](https://mlsanity.com/paper/2408.10823/) (2024)

## Abstract (arXiv)

End-to-end learned lossy image coders (LICs), as opposed to hand-crafted image codecs, have shown increasing superiority in terms of the rate-distortion performance. However, they are mainly treated as black-box systems and their interpretability is not well studied. In this paper, we show that LICs learn a set of basis functions to transform input image for its compact representation in the latent space, as analogous to the orthogonal transforms used in image coding standards. Our analysis provides insights to help understand how learned image coders work and could benefit future design and development.