I create physiological sensing and imaging systems that are based on novel sensor configurations and specifically engineered algorithms that can extract and utilize information hidden within those sensors’ signals. My overall vision is to make people well-connected with digital information on their wellness.
@ Quanttus
Wearable Intelligence for Cardiovascular Health
An MIT-spun out company based out of Cambridge, MA. We focused on tackling the problem of measuring blood pressure using wearable sensing devices. I led the Algorithms development team where we investigated heart rate, respiration, pulse transit time and blood pressure measurement.
@ Microsoft Research
Sensor technology and algorithm development for human health and wellness.
I have developed a robust system for measuring heart rate from conventional RGB cameras. This system is capable of performing robustly while the user performs daily activities such as exercise, desk work or sitting on the couch, watching TV.
Submitted to ACM CHI 2014. More information available soon.
I also enjoy playing with noisy and squiggly lines that represent bio-signals captured from wearable and ambient sensors.
Submitted to ACM CHI 2014. More information available soon.
I also enjoy playing with noisy and squiggly lines that represent bio-signals captured from wearable and ambient sensors.
@ MIT Media Lab (2011 - 2012)
Re-purposing available mobile devices and conventional cameras for medical imaging.
Built a low-cost standalone wearable system for imaging the human retina called PRISM. It exploits the transparent nature of retinal tissues in addition to translucent properties of the skin to illuminate the retina the temporal side of the eye.
Video: Link
Video: Link
Sports Imaging Analytics
Umpires in the game of Cricket and baseball face a tough job when they have to make the call whether a ball hit the bat or not. I investigated some of the imaging technologies in use today to assist the umpires and noticed, well, it is not too hard to fool them.
Video: Link
Video: Link
@ Rochester Institute of Technology (2009 - 2013)
A Better Looking Brain (Ph.D. Dissertation)
I have developed signal processing methods for functional brain imaging by applying two concepts – Wavelets and Independent Component Analysis (ICA). Both these techniques have been used to decompose a signal into its constituent sub-signals; I have extended them to formulate new methods for denoising volumetric images, normalizing multi-subject fMRI (functional magnetic resonance imaging), and synthesizing spatial templates of brain networks.
A perspective on fMRI image processing (included in my thesis): PDF
A perspective on fMRI image processing (included in my thesis): PDF
