Some Latent Representations in Medical Image Analysis and Computer Vision

Tuesday, February 27, 2018 -
4:00pm to 5:00pm
Rm. 1360 Biotechnology Center, 425 Henry Mall

Speaker Name: 

Seong Jae Hwang

Speaker Institution: 

Computer Sciences, UW-Madison




Capturing the underlying attributes of the data with their latent representations (i.e., PCA) often benefits various downstream tasks such as statistical test and classification in many domains. At the same time, the process of identifying the latent representations appropriate to the data and/or analysis is itself an important task. In this talk, I will present some examples of latent representations with different purposes and demonstrate how they tackle the following two problems from very distinct domains: (1) brain connectivity analyses of Alzheimer’s disease in medical imaging and (2) visual relationship learning in computer vision.