Research at a Glance
Computer Vision, Image Processing and Machine Learning
Advanced Educational Strategies
Academia stands at a turning point. The rise of video-based learning and the rapid development of large language models have unsettled long-held assumptions about the value of a traditional university education in the public imagination. As knowledge becomes instantly accessible and instruction increasingly asynchronous, universities must reconsider what makes higher education distinctive. What strategies allow institutions to remain relevant in an era of algorithmic assistance? Which teaching methods genuinely resonate with the next generation of learners? And how can artificial intelligence be integrated into the classroom in ways that enhance thinking rather than create cognitive debt? These questions define a broad and urgent research agenda. Supported by the National Science Foundation and the California Learning Lab
Plant Disease Analysis
Some plant diseases are so destructive that the confirmed presence of a single pathogen can justify removing an entire field. Phytoplasmas and certain bacterial infections produce distinct visual symptoms across leaves and stems, often before yield loss becomes irreversible. Advances in computer vision now make it possible to detect these patterns algorithmically, enabling earlier diagnosis, faster intervention, and reduced economic loss. By integrating AI-assisted image analysis into agricultural practice, disease surveillance can become both more responsive and more cost-effective.
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).
i
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.