Author

Yanru Liu

Date of Award

2025

First Advisor

Kenneth Knox

Second Advisor

Amanda Landi

Abstract

Multiphoton microscopy (MPM) is a useful tool for imaging live, deep, and label free tissues. But, using point scanning in most MPM platforms causes a balance issue between time, field of view (FOV), phototoxicity, and image quality. This often causes noisy images, especially when quick, large FOV, deep, or non-invasive imaging is needed. While deep learning has the potential to denoise MPM data, there is still risk of hallucinations in these algorithms, posing significant challenges for medical and scientific applications. In this project, we introduce a dataset de- signed to facilitate deep learning-based denoising, thereby enhancing the reliability of algorithms for predictions based on deep learning. Additionally, by employ ing this dataset to fine tune a neural network, we demonstrate the preservation of fine features and superior performance compared to other denoising methods.

Simon's Rock Off-campus Download

Simon's Rock students and employees can log in from off-campus by clicking on the Off-campus Download button and entering their Simon's Rock username and password.

Share

COinS