My research interests are in the area of image processing, with applications in medical imaging and multimedia. During the past one to two decades, there has been a significant increase in the level of interest in designing efficient image analysis systems for multimedia and medical applications. Many new application areas, such as the computer aided diagnostic systems, image visualization, content-based image and video retrieval, multimedia communications, image and video databases are feasible with the current technology. However, as new applications are coming up, new requirements are being specified, and a substantial amount of work remains to be accomplished in this area. My long term interest is in the development of image analysis techniques that help in computer aided diagnosis of selected critical diseases as well as in multimedia data analysis.
Over the past years my work focused on image and video compression, content-based image retrieval, and wavelet analysis. Currently, my research focus is primarily in medical imaging and super-resolution imaging.
Computer-aided-diagnosis: We are developing efficient image analysis algorithms for healthcare applications. Three applications are being considered: (i) radiograph image analysis to detect TB; (ii) histopathological image analysis to detect melanoma (one type of skin cancer); and (iii) capsule endoscopic image analysis to detect bleeding in the gastro-intestinal tract. The developed techniques will be helpful in automated detection of critical diseases and also can be useful as a second opinion.
Image analysis and feature detection using stochastic resonance: We are developing efficient image analysis techniques using stochastic resonance phenomenon. Both theoretical development of stochastic resonance and its application in signal and image processing are being considered. These techniques will be very useful for image analysis in a noisy environment. Two applications are considered: detection of lesions/tumors in MRI brain images (health application), and digital watermarking (consumer imaging application).
Super resolution imaging and visualization: Here, our focus is on designing efficient low complexity techniques for generating high resolution videos by combining information from several low resolution videos. The developed techniques will be useful in areas such as medical data visualization, sports training, event modeling, and high definition video generation.