We investigate a way to detect images of surface plasmon scattering using deep learning approach. Unlike fluorescence imaging, the image of surface plasmon scattering shows much worse resolution due to propagation length of surface plasmon polariton. In this work, deep learning approach is taken to address this issue and to discriminate multiple target objects under complex and noisy environment. Conventional detection method based on fourier filtering and deconvolution was employed to compare the performance of the proposed method. It was shown that deep learning improves the accuracy by about six times, and especially more useful in noisy environment.

Date Published: 10 March 2020
PDF
Proc. SPIE 11257, Plasmonics in Biology and Medicine XVII, 112570N (10 March 2020); doi: 10.1117/12.2544169

Published in SPIE Proceedings Vol. 11257:
Plasmonics in Biology and Medicine XVII
Tuan Vo-Dinh; Ho-Pui A. Ho; Krishanu Ray, Editor(s)