Computational pathology has seen rapid growth, driven by advanced deep-learning algorithms. However, there's a lack of open-source software providing a generic end-to-end API for pathology image analysis. This paper presents TIAToolbox, a Python toolbox designed to make computational pathology accessible. TIAToolbox creates modular and configurable components for common sub-tasks (WSI data reading, patch extraction, stain normalization, model inference, visualization). The authors demonstrate its use in constructing a deep-learning pipeline and reimplementing state-of-the-art algorithms.

Johnathan Pocock, Simon Graham, Quoc Dang Vu, Mostafa Jahanifar, Srijay Deshpande, Giorgos Hadjigeorghiou, Adam Shephard, Raja Muhammad Saad Bashir, Mohsin Bilal, Wenqi Lu, David Epstein, Fayyaz Minhas, Nasir M. Rajpoot, Shan E Ahmed Raza