arXiv Open Access 2023

Pushing The Limits of the Wiener Filter in Image Denoising

Clément Bled François Pitié

Lihat Sumber

Abstrak

As modern image denoiser networks have grown in size, their reported performance in popular real noise benchmarks such as DND and SIDD have now long outperformed classic non-deep learning denoisers such as Wiener and Wavelet-based methods. In this paper, we propose to revisit the Wiener filter and re-assess its potential performance. We show that carefully considering the implementation of the Wiener filter can yield similar performance to popular networks such as DnCNN.

Topik & Kata Kunci

eess.IV

Penulis (2)

Clément Bled

François Pitié

Format Sitasi

APA MLA BibTeX

Bled, C., Pitié, F. (2023). Pushing The Limits of the Wiener Filter in Image Denoising. https://arxiv.org/abs/2303.16640

Akses Cepat

Lihat di Sumber

Informasi Jurnal

Tahun Terbit: 2023
Bahasa: en
Sumber Database: arXiv
Akses: Open Access ✓