Batch Denoising via Blahut-Arimoto

Qing Li, Cyril Guyot

Abstract:

In this work, we propose to solve batch denoising using Blahut-Arimoto algorithm (BA).
Batch denoising via BA (BDBA), similar to Deep Image Prior (DIP), is based
on an untrained score-based generative model. Theoretical results show that our
denoising estimation is highly likely to be close to the best result. Experimentally,
we show that BDBA outperforms DIP significantly.

Student Paper: No