Video Forthcoming 

Dissertation Title: Autonomic Activity from Human Videos 

Zoom link: https://mit.zoom.us/j/98100236764?pwd=bzdDSkxDVlZSMFk5cC84WDFKVzRxdz09

Abstract: 

Autonomic nervous system (ANS) is part of the nervous system that is responsible for regulation and integration of internal organs’ functioning. Traditionally, for the assessment of the ANS function, autonomic activity is measured in various tests by medical devices with contact sensors. Most of these tests require wearing cumbersome equipment on the human body, so they are commonly conducted in clinics and only short-term data can be collected.

A potential solution to long-term measurement of autonomic activity is via camera-based human sensing. Recent research has shown that they can be combined with computer vision algorithms to realize non-contact estimation of ANS activity parameters such as heart rate, respiration rate, and heart rate variability (HRV). However, there are still many hurdles that prevent the solution from reaching the accuracy and covering the scope of clinical tests: 1) The robustness of the existing methods are still unsatisfactory in ambulatory situations, especially when specular reflection and non-rigid body motions are significant. 2) Compared with heart rate and respiration rate, the measurement of HRV has much lower accuracy and much higher sensitivity to motion artifacts.

To address these problems, this dissertation proposes an end-to-end convolutional attention network using both gradient descent and gradient ascent to enable robust measurement and enhancement under heterogeneous lighting and major motions, proposes a learning-based HRV estimator that can recover inter-beat intervals from noisy ambulatory data, and proposes a framework exploring, revealing, and verifying autonomic activity in unlabeled human video by integrating estimation and visualization. Through combining these proposed approaches, the final goal of the dissertation is to realize unobtrusive analysis of autonomic activity from human video that can work in the field.

Committee members: 

Rosalind W. Picard
Professor of Media Arts and Sciences
Director of the Affective Computing Research Group
Massachusetts Institute of Technology

William T. Freeman
Thomas and Gerd Perkins Professor of Electrical Engineering and Computer Science
Massachusetts Institute of Technology

Ramesh Raskar
Associate Professor of Media Arts and Sciences
NEC Career Development Professor
Massachusetts Institute of Technology