Showing 1–1 of 1 results for author: Zambrano, R
-
Self-supervised OCT Image Denoising with Slice-to-Slice Registration and Reconstruction
Authors:
Shijie Li,
Palaiologos Alexopoulos,
Anse Vellappally,
Ronald Zambrano,
Wollstein Gadi,
Guido Gerig
Abstract:
Strong speckle noise is inherent to optical coherence tomography (OCT) imaging and represents a significant obstacle for accurate quantitative analysis of retinal structures which is key for advances in clinical diagnosis and monitoring of disease. Learning-based self-supervised methods for structure-preserving noise reduction have demonstrated superior performance over traditional methods but fac…
▽ More
Strong speckle noise is inherent to optical coherence tomography (OCT) imaging and represents a significant obstacle for accurate quantitative analysis of retinal structures which is key for advances in clinical diagnosis and monitoring of disease. Learning-based self-supervised methods for structure-preserving noise reduction have demonstrated superior performance over traditional methods but face unique challenges in OCT imaging. The high correlation of voxels generated by coherent A-scan beams undermines the efficacy of self-supervised learning methods as it violates the assumption of independent pixel noise. We conduct experiments demonstrating limitations of existing models due to this independence assumption. We then introduce a new end-to-end self-supervised learning framework specifically tailored for OCT image denoising, integrating slice-by-slice training and registration modules into one network. An extensive ablation study is conducted for the proposed approach. Comparison to previously published self-supervised denoising models demonstrates improved performance of the proposed framework, potentially serving as a preprocessing step towards superior segmentation performance and quantitative analysis.
△ Less
Submitted 8 December, 2023; v1 submitted 25 November, 2023;
originally announced November 2023.
