Cancer cytoplasm segmentation in hyperspectral cell image with data augmentation

TL;DR

This paper presents a deep learning approach for cancer cytoplasm segmentation in hyperspectral cell images, utilizing a novel data augmentation technique based on CMOS images to improve model performance amidst limited data and noise.

Contribution

The study introduces a data augmentation method using CMOS images to enhance deep learning-based cancer cytoplasm segmentation in hyperspectral images, addressing data scarcity and noise issues.

Findings

Data augmentation improved segmentation accuracy.

The method effectively handled instrumental noise.

Enhanced model robustness demonstrated experimentally.

Abstract

Hematoxylin and Eosin (H&E)-stained images are commonly used to detect nuclear or cancerous regions in cells from images captured by a microscope. Identifying cancer cytoplasm is crucial for determining the type of cancer; hence, obtaining accurate cancer cytoplasm regions in cell images is important. While CMOS images often lack detailed information necessary for diagnosis, hyperspectral images provide more comprehensive cell information. Using a deep learning model, we propose a method for detecting cancer cell cytoplasm in hyperspectral images. Deep learning models require large datasets for learning; however, capturing a large number of hyperspectral images is difficult. Additionally, hyperspectral images frequently contain instrumental noise, depending on the characteristics of the imaging devices. We propose a data augmentation method to account for instrumental noise. CMOS images were used for data augmentation owing to their visual clarity, which facilitates manual annotation compared to original hyperspectral images. Experimental results demonstrate the effectiveness of the proposed data augmentation method both quantitatively and qualitatively.