Sub2Full: split spectrum to boost OCT despeckling without clean data
This addresses the problem of image quality deterioration in OCT, especially for high-resolution modalities like visible light OCT, by providing a practical solution without requiring clean data, though it is incremental as it builds on self-supervised denoising methods.
The paper tackles speckle noise in optical coherence tomography (OCT) images, which degrades quality, by proposing Sub2Full, a self-supervised method that avoids the need for clean data and demonstrates superior performance over existing schemes like Noise2Noise and Noise2Void.
Optical coherence tomography (OCT) suffers from speckle noise, causing the deterioration of image quality, especially in high-resolution modalities like visible light OCT (vis-OCT). The potential of conventional supervised deep learning denoising methods is limited by the difficulty of obtaining clean data. Here, we proposed an innovative self-supervised strategy called Sub2Full (S2F) for OCT despeckling without clean data. This approach works by acquiring two repeated B-scans, splitting the spectrum of the first repeat as a low-resolution input, and utilizing the full spectrum of the second repeat as the high-resolution target. The proposed method was validated on vis-OCT retinal images visualizing sublaminar structures in outer retina and demonstrated superior performance over conventional Noise2Noise and Noise2Void schemes. The code is available at https://github.com/PittOCT/Sub2Full-OCT-Denoising.