PatchDenoiser: Parameter-efficient multi-scale patch learning and fusion denoiser for medical images

Jitindra Fartiyal, Pedro Freire, Sergei K. Turitsyn, Sergei G. Solovski · 2026

Read PatchDenoiser: Parameter-efficient multi-scale patch learning and fusion denoiser for medical images by Jitindra Fartiyal, Pedro Freire, Sergei K. Turitsyn, Sergei G. Solovski (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-21987.txt.

Description

Low-dose CT scans cut radiation risk for children, cancer screening, and repeat checkups, but the images come out noisy. Jitindra Fartiyal and three co-authors note that old filters blur away fine detail, while heavy deep-learning denoisers need large, costly models. They propose PatchDenoiser, a light multi-scale method that splits the job into local texture work and broad context work, fused patch by patch with attention to place.

The design quiets noise while keeping fine body structures intact. It does so with far fewer parameters and far less computing than rival methods based on convolutional nets, adversarial nets, or transformers. Tests on standard low-dose data show it holds its own while staying cheap to run. For clinics, that mix of clarity and low cost could make low-dose scanning more practical day to day.

Catalog record

CREATORJitindra Fartiyal, Pedro Freire, Sergei K. Turitsyn, Sergei G. Solovski
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE33 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-21987

Browse the full library · About the Tower of Babel