# Autoencoder based image compression: can the learning be quantization independent?

arXiv: https://arxiv.org/abs/1802.09371 · HTML: https://mlsanity.com/paper/1802.09371/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Feb 2018.

## Built on by (most cited first, 8 of 8)

- [RGB no more: Minimally-decoded JPEG Vision Transformers](https://mlsanity.com/paper/2211.16421/) (2022)
- [Learning-Driven Lossy Image Compression; A Comprehensive Survey](https://mlsanity.com/paper/2201.09240/) (2022)
- [HCFL: A High Compression Approach for Communication-Efficient Federated Learning in Very Large Scale IoT Networks](https://mlsanity.com/paper/2204.06760/) (2022)
- [Make Lossy Compression Meaningful for Low-Light Images](https://mlsanity.com/paper/2305.15030/) (2023)
- [Extreme-scale Talking-Face Video Upsampling with Audio-Visual Priors](https://mlsanity.com/paper/2208.08118/) (2022)
- [Image Storage on Synthetic DNA Using Autoencoders](https://mlsanity.com/paper/2203.09981/) (2022)
- [Image storage on synthetic DNA using compressive autoencoders and DNA-adapted entropy coders](https://mlsanity.com/paper/2306.12882/) (2023)
- [Powerful Lossy Compression for Noisy Images](https://mlsanity.com/paper/2403.14135/) (2024)

## Abstract (arXiv)

This paper explores the problem of learning transforms for image compression via autoencoders. Usually, the rate-distortion performances of image compression are tuned by varying the quantization step size. In the case of autoen-coders, this in principle would require learning one transform per rate-distortion point at a given quantization step size. Here, we show that comparable performances can be obtained with a unique learned transform. The different rate-distortion points are then reached by varying the quantization step size at test time. This approach saves a lot of training time.