Research at a Glance
Computer Vision, Image Processing and Machine Learning
Affect Analysis and Recognition
Up to 55% of communication occurs via non-verbal cues. Computer vision systems have the potential to automatically detect apparent facial expression to determine underlying affect. This has wide-spread applications in lie detection, human-computer interaction and observation analysis (pictured, collaboration with Motor Trend).
Videobioinformatics
Despite the tremendous advances in automatic computer analysis, a significant proportion of video-based biological research is still quantified by time-consuming and subjective human analysis. Computer vision systems can be applied to quickly analyze videos enabling a high-thoroughput and accuracy. Featured in Science.
Bio-inspired Computer Vision
One of the great challenges to artificial intelligence is the emulation and understanding human behavior. While the human brain holds a great many mysteries, vision and face processing in particular are among the most well understood cognitive processes. In bio-inspired computer vision, the human visual system is used as inspiration for vision algorithms.
Thermographic Analysis of Photovoltaics
By 2050, the world will need an estimated 20 terawatts of non-CO2 energy to curb the impending greenhouse gas production crisis. Vision technologies can be deployed in parallel with these systems for defect detection, quality control and monitoring of systems. Advanced computer methods are being developed to allow utilities to remotely assess the health of PV panels at quickly at a large-scale with low-cost IR vision systems.
