Sandhitsu Das

biography

After finishing my undergraduate training, I got my Masters in Electrical Engineering specializing in digital signal processing -- both from the Indian Institute of Technology, Kanpur. During my final years in Kanpur, I worked on representation of music in the brain and a teleradiology kit for brain segmentation, under the guidance of the inspirational Dr. G.C. Ray. I became enamored by the intricacies of the human brain -- and decided to join the Neuroengineering Laboratory at the University of Pennsylvania to work with Dr. Leif Finkel.

My doctoral research focused on the human brain's intriguing ability to perceive humans and animals in motion even in the presence of sparse visual information -- and its application in computer vision. In my research, I analyzed human motion data, and used psychophysical experiments to make a humble attempt to say something about what the brain might be doing while it does biological motion recognition. Based on what we might learn about how the brain processes human motion, we proposed a technique for gait recognition from motion capture data based on two successive stages of principal component analysis (PCA) on kinematic data. Components of this representation closely correspond to particular spatiotemporal features of gait that we have shown to be important for visual recognition of gait. The projection space of PCA has distinguishable clusters corresponding to the individual identity and type of gait.

After finishing my Ph.D., I shifted my focus to neuroimaging. I joined PICSL in 2006 as a postdoctoral researcher. In particular, I'm interested in applying image analysis technologies to study pathologies in the human brain, as well as its functional organization. My current research includes:

• I'm developing image-based tools for measuring cortical thickness which is a valuable clinical measure for diagnosis and evaluation of neurodegenrative diseases.
• I'm developing tools for automated labeling of visual areas for retinotopic mapping. This includes cortical surface flattening and atlas fitting procedures.
• I'm analyzing functional MRI data from epilepsy patients to quantify asymmetric activations of hippocampi using geometric model-based normalization.

bibliography

Journal Papers

Peer-Reviewed Conference Papers

Conference Abstracts