Presentation
7 March 2022 Polarimentric differentiation of ex vivo colon samples complemented by machine-learning
Author Affiliations +
Abstract
We present a combination of Mueller matrix measurements (635 nm) of cancerous colon specimens and machine-learning approach. Physical realizability filtering and symmetric decomposition were used to extract polarimetric quantities, used as predictors in machine-learning algorithms. The results were visualized using various depolarization spaces. Principal component analysis was used to extract particular features from the model, logistic regression evaluated predictors with high likelihood for tumor detection, while random forest and support vector machines provided the best results for classification. Hence, polarimetry combined with machine-learning approach may increase the histopathology diagnostic accuracy.
Conference Presentation
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Deyan Ivanov, Viktor Dremin, Tsanislava Genova, Alexander Bykov, Igor Meglinski, Tatiana Novikova, and Razvigor Ossikovski "Polarimentric differentiation of ex vivo colon samples complemented by machine-learning", Proc. SPIE PC11963, Polarized Light and Optical Angular Momentum for Biomedical Diagnostics 2022, PC1196307 (7 March 2022); https://doi.org/10.1117/12.2615297
Advertisement
Advertisement
KEYWORDS
Polarimetry

Colon

Tissue optics

Diagnostics

Optical spheres

Polarization

Principal component analysis

Back to Top