My research interests concentrate on next-generation visual information processing, and more specifically on:

• digital image and video processing, compression and transmission,
• visual sensor networks (camera webs),
• stereoscopic and 3-D imaging,
• multidimensional digital signal processing.

Currently, my main focus is on applications in the following areas:

• visual surveillance, especially video analytics and video forensics (homeland security, including mass-transport and airports, healthcare facilities),
• environmental monitoring (pervasive, nonintrusive, autonomous observation of natural habitat of various species),
• 3-D TV and 3-D multimedia (stereoscopic and multi-view processing and compression for next-generation services).

The research is carried by graduate and undergraduate students within the Visual Information Processing (VIP) Laboratory. Currently, the following main research thrusts are pursued:

• Visual Sensor Networks (VSNs): Statistical Models and Methods for Video Analytics in Camera Networks
  Visual sensor networks (or camera webs) are a reality today; over 30 million surveillance cameras in the US alone generate more than 4 billion hours of video footage each week. Our focus is on the development of advanced algorithms for video analytics in order to enable pervasive, autonomous (no human operator in the loop) and rate-efficient visual surveillance.
   
• Video: Next-Generation Methods for Image Sequence Analysis, Processing, and Transmission
  This research concentrates around the idea that by processing multiple frames of an image sequence jointly, one can achieve not only better results, e.g., more accurate segmentation or improved compression, but also extract qualitatively new information, such as occlusion and newly-exposed area information.
   
• 3-D: Image Processing and Compression Framework 3-D Visual Communications
  The focus here is on multiview acquisition, compression and enhancement algorithms for automultiscopic 3-D displays, i.e., displays that do not require glasses while providing multiple perspectives to the viewer (look-around). Such displays have recently been introduced on the market and hold a great promise for the future of 3-D visual communications.
   
• BioMed: Processing and Visualization of Biomedical Imagery
  3-D visualization is an indispensable tool in the analysis of biomedical data. Moreover, in some scenarios such visualization must not use shutter glasses for their visual interference with neighboring 2-D monitors. It is the purpose of this research axis to explore the potential of automultiscopic displays in scientific, and especially biological and medical, applications.