A framework for building tissue classification pipelines of digital microscopy images: a case study in breast cancer

📄 Project Abstract

Histopathology is the study of disease via the analysis of tissue at the microscopic level. Histopathologists manually analyse slides of tissue in order to detect the presence of disease. One area in which this is done is the analysis of breast tissue in order to detect breast cancer. Automated tissue analysis using computer vision methods is the focus of much past and present research. The aim of such research is to develop systems to automate the process of tissue analysis and take some of the workload off the Histopathologists. This has been made possible by advances in the process of slide digitisation opening up many exciting research opportunities. This study hopes to contribute to this research by presenting and evaluating a framework for building systems to automatically classify areas of skin tissue, which, it is hoped will be of use to future research in this area. These areas of tissue are defined by `spots', locations of small areas in the tissue that are the focus for clas...