Imaging Science Seminar - Small Data, Big Insights

Small Data, Big InsightsDr. Raja BalaPrincipal Scientist, Computer Vision GroupPalo Alto Research CenterAbstract:Deep learning has defined the state of the art for many computer vision tasks, thanks to advances in computing hardware and the availability of large datasets. However, for many practical applications, data acquisition and annotation can be a laborious and expensive task. This is especially true in specialized fields such as medical imaging, where annotation must be carried out by domain experts. I will describe a series of novel approaches developed at PARC to address the “Small Data” problem. This includes i) incorporating domain-specific regularizers into deep networks for robustness to sparse training; ii) data augmentation via regularized generative models; iii) automatic image annotation based on pre-trained model ensembles; iv) interactive scene annotation with augmented reality; and v) bringing the human into the loop for high quality image acquisition.Speaker Bio:Raja Bala is a Principal Scientist and Leader of the Computer Vision group at Palo Alto Research Center. His research interests are at the intersection of digital imaging, computer vision, and machine learning. Earlier he was a Principal Engineer at Samsung Research America, where he developed innovative computational imaging technologies for Samsung’s high-end smartphones. Previous to that, Bala worked for Xerox and PARC, leading efforts in a variety of imaging and vision applications, including facial health analysis, mobile document imaging, human activity monitoring with wearable sensors, license plate recognition, and traffic anomaly detection. He has served as adjunct faculty in the School of Electrical Engineering at the Rochester Institute of Technology. He is a Fellow of IS&T and Senior Member of IEEE, and holds over 100 publications and 170 patents. He is Associate Editor of the IEEE Transactions on Image Processing, and co-editor of the book “Computer Vision and Imaging in Intelligent Transportation Systems”. He received a Ph.D. in Electrical Engineering from Purdue University.