An ultrathin fiber-based fluorescent imaging probe based on hyperspectral imaging

Frontiers in Physics

Physics

Shuyan Zhang¹  Malini Olivo¹  Xiuting Li¹  Joseph Yong Xin Cheng²  Jian Jun Chua²  Wilbert Yao Hwee Tang² 

[1] Translational Biophotonics Laboratory, Institute of Bioengineering and Bioimaging, A*STAR, Singapore, Singapore;Translational Biophotonics Laboratory, Institute of Bioengineering and Bioimaging, A*STAR, Singapore, Singapore;Department of Biomedical Engineering, National University of Singapore, Singapore, Singapore;

关键词: fiber sensor;    hyperspectral imaging;    fluorescent imaging;    spectral unmixing;    bioimaging;    quantum dots;    bacteria;    spectral angle mapper;   

DOI  :  10.3389/fphy.2022.1096290

 received in 2022-11-11, accepted in 2022-12-12,  发布年份 2023

来源: Frontiers

【 摘 要 】

Most fluorescent imaging systems use pre-defined filter sets with limited flexibility in the choice of excitation and emission bands and suffer from crosstalk between different fluorophores for multiplexed imaging. When implemented in the context of optical fibers for biological imaging in a remote setting with space constraints, the size of the fiber probe is crucial. To overcome these challenges, we combined the merits of hyperspectral imaging and an ultrathin optical imaging fiber where fluorescent images at 311 spectral bands in the visible range were captured using a liquid crystal tunable filter with a fiber probe of 500 μm in diameter. Fluorescent experiments were performed on quantum dots mixtures and genetically modified E. coli bacteria mixtures to demonstrate the system’s capabilities for multiplexed imaging. Images were first processed using a Fourier transform filtering technique to remove the fiber core pattern artifact. Subsequently, the classification of different types of bacteria samples was calculated using two methods, namely, spectral unmixing with maximum abundance and spectral matching with minimum spectral angle distance. Finally, the spatial distributions of individual bacteria types were overlaid with the mixture image, and the two classification results matched well. A simple-to-use graphic user interface (GUI) platform for the hyperspectral imaging fiber probe system was developed which performs image processing and displays the classification results. The methods and results presented will appeal to the optical fiber-based imaging and hyperspectral imaging communities in general and have great potential for biological imaging applications.

【 授权许可】

Unknown   
Copyright © 2023 Zhang, Chua, Tang, Cheng, Li and Olivo.