BROWSE

Related Researcher

Author's Photo

Chun, Se Young
Bio-Medical Image Processing Lab (BMIPL)
Research Interests
  • Inverse problem, sparse signal, multimodal information, diffeomorphic alignment, statistical learning, medical imaging

ITEM VIEW & DOWNLOAD

On Divergence Approximations for Unsupervised Training of Deep Denoisers Based on Stein’s Unbiased Risk Estimator

Cited 0 times inthomson ciCited 0 times inthomson ci
Title
On Divergence Approximations for Unsupervised Training of Deep Denoisers Based on Stein’s Unbiased Risk Estimator
Author
Soltanayev, ShakarimGiryes, RajaChun, Se YoungEldar, Yonina C.
Issue Date
2020-05-04
Publisher
IEEE
Citation
IEEE International Conference on Acoustics, Speech and Signal Processing, pp.3592 - 3596
Abstract
Recently, there have been several works on unsupervised learning for training deep learning based denoisers without clean images. Approaches based on Stein’s unbiased risk estimator (SURE) have shown promising results for training Gaussian deep denoisers. However, their performance is sensitive to hyper-parameter selection in approximating the divergence term in the SURE expression. In this work, we briefly study the computational efficiency of Monte-Carlo (MC) divergence approximation over recently available exact divergence computation using backpropagation. Then, we investigate the relationship between smoothness of nonlinear activation functions in deep denoisers and robust divergence term approximations. Lastly, we propose a new divergence term that does not contain hyper-parameters. Both unsupervised training methods yield comparable performance to supervised training methods with ground truth for denoising on various datasets. While the former method still requires roughly tuned hyper parameter selection, the latter method removes the necessity of choosing one.
URI
https://scholarworks.unist.ac.kr/handle/201301/32634
URL
https://ieeexplore.ieee.org/document/9054593
DOI
10.1109/icassp40776.2020.9054593
Appears in Collections:
AI_Conference Papers
Files in This Item:
There are no files associated with this item.

find_unist can give you direct access to the published full text of this article. (UNISTARs only)

Show full item record

qrcode

  • mendeley

    citeulike

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

MENU