COIN: COmpression with Implicit Neural representations

By Emilien Dupont, Adam Golinski, Milad Alizadeh, Yee Whye Teh, Arnaud Doucet. International Conference on Learning Representations, 2021

332 citations recorded by the source corpus.

We propose a new simple approach for image compression: instead of storing the RGB values for each pixel of an image, we store the weights of a neural network overfitted to the image. Specifically, to encode an image, we fit it with an MLP which maps pixel locations to RGB values. We then quantize and store the weights of this MLP as a code for the image. To decode the image, we simply evaluate the MLP at every pixel location. We found that this simple approach outperforms JPEG at low bit-rates, even without entropy coding or learning a distribution over weights. While our framework is not yet competitive with state of the art compression methods, we show that it has various attractive properties which could make it a viable alternative to other neural data compression approaches.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/1bf444b861acc3dad72d968c2c69bcb863885ff9

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Ser-Nam Lim
2. Yixuan Ren
3. Abhinav Shrivastava
4. Bo He
5. Hanyu Wang
6. Hao Chen
7. Luc Van Gool
8. Dacheng Tao

Citations in the corpus

• NeRV: Neural Representations for Videos (2021, background): “Most recently, [17] demonstrated the feasibility of using implicit neural representation for image compression tasks.”