SpecEye: A Novel Open-Source Desktop Application for Hyperspectral Fluorescence Microscopy Imaging

Abstract

Purpose : Fluorescence microscopic imaging of tissue sections serves as a foundational tool in diagnostic pathology and biomedical research. The combination of fluorescence microscopy with a hyperspectral imager, such as an image mapping spectrometer (IMS), has enabled the acquisition of datasets with high spectral and spatial resolutions. There is an unmet need for efficient image processing techniques for diverse types of FMHSI. Our aim is to develop an innovative and open-source desktop platform using Python, accessible across Mac, Windows, and Linux operating systems for the preprocessing, visualization, semantic segmentation and boundary detection of FMHSI, with a special focus on eye tissue sections

Methods : This open-source platform, "SpecEye", was developed based on HFMI datacubes obtained from frozen eye sections of C57BL/6 and B6 Albino mice imaged with an IMS imager mounted on a fluorescence microscope. The main objective of our study is to develop segmentation and boundary detection algorithms to investigate the distribution of biomolecules with intrinsic fluorescence in the layer-specific eye tissues. Those algorithms include spectral information divergence spectral angle mapper with optional unmixing and spatial fuzzy C-means clustering with Sobel edge detector using spatial and spectral information from label free FMHSI eye images. This proposed app offers a suit of functionalities: data preprocessing tools such as normalization, denoising and superpixel generation; visualization tools such as 2D and 3D spectral-based interactive exploration, region of interest selection, and initial identification of endmember signatures and the segmentation tools.

Results : SpecEye offers a user-friendly interface for analyzing FMHSI across various environments, equipped with a diverse set of tools tailored to different objectives. Moreover, it streamlines the unsupervised label-free segmentation of eye layers to reduce the time by an expert.

Conclusions : The implementation of SpecEye broadens the scope of FMHSI capacities, empowering the users to gain deeper insights into the intricacies of hyperspectral biomedical data. Moreover, this integrated application holds significant promise to validate data obtained with state of art in vivo eye imaging with tissue sections imaged with hyperspectral fluorescence microscopy. Also, this app can be applied for FMHSI data from other studies.

This abstract was presented at the 2024 ARVO Imaging in the Eye Conference, held in Seattle, WA, May 4, 2024.