Explore the intersection of deep learning and biomedicine in this 25-minute conference talk by Dr. Olivier Gevaert at Devoxx. Delve into the challenges and opportunities presented by massive amounts of biomedical data, ranging from molecular to tissue-scale. Learn about the concept of precision medicine and how deep learning techniques are being applied to analyze complex, multi-modal biomedical datasets. Discover the types of data used in biomedicine, including images, text, and numeric data, and understand the potential applications and challenges in modeling this information using deep learning approaches. Gain insights into multi-scale biomedical data fusion, with a focus on oncology applications, and explore how machine learning methods are being developed to find relevant patterns in high-dimensional, multi-modal datasets. Follow Dr. Gevaert's expertise in combining elements from statistics and mathematics to analyze data from molecular biology, pathology, and radiology.
Deep Learning in Biomedicine - Challenges and Opportunities
Overview
Syllabus
Intro
Precisions medicine
Big data vs. data science
Precision medicine
Biomedical data
Fast technological development
The Cancer Genome Atlas (TCGA)
Digital pathology
Biomedical images
Medical imaging
Introduction: Radiomics
Introduction: Traditional Radiomics
Semantic Features: Edge shape
Computational Features: Overview
Stanford Data Set
Decision tree modeling
Classifier: Decision trees
Introduction: Deep learning
Deep learning for medicine
Convolutional Neural Network
Melanoma diagnosis
Different types of images
Lessons learned
Background and Motivation
Conventional Approach
Brain Tumor Segmentation (Brats) Challenge
Ground truth
Visualizations
Overall summary
Taught by
Devoxx
