Accurate and Interpretable Classification of Microspectroscopy Pixels Using Artificial Neural Networks

Published

January 6, 2017

Author(s)

Petru S. Manescu, Young Jong Lee, Charles Camp, Marcus T. Cicerone, Mary C. Brady, Peter Bajcsy

Abstract

This paper addresses the problem of classifying materials from microspectroscopical images at each pixel. The challenges lie in identifying discriminatory spectral features and obtaining accurate and interpretable models relating spectra and class labels. We approach the problem by designing a supervised classifier from a tandem of Artificial Neural Network (ANN) models that identify relevant features in raw spectra and achieve high classification accuracy. The tandem of ANNs models is meshed with classification rule extraction methods to lower the complexity and to achieve interpretability of the resulting model. The novelty of the work is in designing each ANN model based on the microspectroscopical imaging hypothesis about a target class being related to a linear combination of spectra, and in meshing ANN and decision rule models into a tandem configuration to achieve accurate and interpretable classification results. The proposed method was evaluated using a set of broadband coherent anti-Stokes Raman scattering (BCARS) microscopy cell images (600,000 pixel-level spectra) and a reference four-class rule-based model previously created by biochemical experts. The generated classification rule-based model was on average 85% accurate measured by the DICE index over all pixels, and on average 96% similar to the reference rules measured by the vector cosine metric.

Citation

Medical Image Analysis

Volume

37

Pub Type

Journals

Download Paper

Local Download

Keywords

Microspectroscopy, Artificial Neural Networks, BCARS, hyperspectral imaging, rule-based model

Citation

Manescu, P. , Lee, Y. , Camp, C. , Cicerone, M. , Brady, M. and Bajcsy, P. (2017), Accurate and Interpretable Classification of Microspectroscopy Pixels Using Artificial Neural Networks, Medical Image Analysis, [online], https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=920912 (Accessed November 21, 2024)