Patch-Based Convolutional Neural Networks for TCGA-BRCA Breast Cancer Classification
Document Type
Conference Proceeding
Publication Date
2021
Abstract
The current study automatically identified regions of interest and classified breast tumors in whole slide images from The Cancer Genome Atlas Breast Invasive Carcinoma (TCGA-BRCA) using patch-based convolutional neural networks (CNNs). Pre-processing techniques were applied on whole slide images. Then, whole slide images were tiled into patches, and patches containing regions of interest (ROIs) like nuclei-rich areas were identified. Afterwards, features from patches containing ROIs were extracted using CNNs and used to train patch-level classifiers. Finally, patch-level predictions were aggregated into slide-level predictions. Classification metrics like accuracy, precision, recall, and f1-score were used to evaluate results.
Recommended Citation
Villareal, R. J. T., & Abu, P. A. R. (2022). Patch-based convolutional neural networks for TCGA-BRCA breast cancer classification. In G. Bebis, V. Athitsos, T. Yan, M. Lau, F. Li, C. Shi, X. Yuan, C. Mousas, & G. Bruder (Eds.), Advances in Visual Computing (pp. 29–40). Springer International Publishing. https://doi.org/10.1007/978-3-030-90436-4_3
