One of the main indicators for cancer diagnosis and assessment is the number of mitosis observed in histology sections. This figure is normally obtained by observation of histopathology preparations by expert histologists. Now well, this is a routine task, that requires a long time and consequently it has a high with no trace of recognition of the mitosis.
Moreover, despite the experience of the pathologists, the manual annotation is not completely objective, the result can be different for two different experts or even different for the same expert at different times. The availability of computational tools that automate this process would avoid most of these issues.
In this course, students will familiarize themselves with the basics of the technologies used to be able to construct this type of tools, and they will have access to its use through a case study with real data.
The field of digital pathology image interpretation and analysis, in general, and the mitosis detection, in particular, involve a large number of concepts and techniques. In this sense, the objective of this course is not so much the comprehensive training of students in this area, as the introduction of the most basic aspects. Thus, a general and practical vision of the whole image analysis process is presented for a particular use case with the aim of providing students with the concepts and tools that facilitate a deeper immersion in this field in a personal way. In addition, we intend to foster students' curiosity and enthusiasm for an area as exciting as image analysis processing and analysis for particular applications.
Módulos del curso
Module 0. Welcome
Module 1. Introduction
Module 2. Introduction to medical imaging
Module 3. Fundamentals of Digital Image Processing
Module 4. Machine Learning
Module 5. Use case: Automatic mitosis detection in histologic images
Module 6. Final Project