The Human Sensing Laboratory (HSL) is located in the Smith Hall building in the Robotics Institute.
We work on three areas:
Enabling computers to understand and characterize human behavior has the potential to revolutionize many areas that benefit society such as clinical diagnosis, human computer interaction, and social robotics. In the Human Sensing (HS) laboratory we are interested in modelling and characterizing human behavior from a variety of sensory data (e.g., video, motion capture, wearable sensors). Current projects include: (1) Depression detection from multimodal data (e.g., RGBD video, binary sensors and accelerometers), (2) monitoring Parkinson's desease with wearable sensors, (3) automatic stress analysis from psychological measurements (e.g., EEG, GSR, heart rate) and (4) hot-flash detection with wearable GSR sensors.
Component Analysis (CA) methods (e.g. kernel PCA, Support Vector Machines, Spectral Clustering) are a set of algebraic techniques that decompose a signal into components that are relevant for a given task (e.g., classification, clustering). In the CA laboratory we are interested in extending CA techniques for classification, clustering, modelling and visualization of large amounts of high-dimensional data. Some of our current work include Time series (supervised, unsupervised, weakly-supervised), Temporal clustering and alignment, Graph Matching (i.e., quadratic assignment problems) and kernel methods.
The face is one of the most powerful channels of non-verbal communication. Several of our current research projects explore the use of facial behaviour as a predictor of internal states, social behaviour, biometrics and psychopathology. Current work includes Facial feature detection and tracking, Facial Expression Analysis, Face Recognition, Facial Expression Transfer, Face De-identification and Facial attribute estimation (e.g., age, ethnicity)