Current Research Interests
A large part of my research involves bringing in (artificial) intelligence into data (particularly, image data) acquisition. In this era of BIG data, one should acquire data smartly and measure only what is really needed for the applications
at hand. This is a huge motivation for compressed sensing, i.e. acquisition of data directly in compressed format for improvement of acquisition speed and reduction of resources for acquisition. Such resources could include time, radiation, battery power, electricity, etc.
Keeping this philsophy in mind, my research interests include the following. A brief presentation regarding some of these topics can be found here.
Take a look at our publications, as well as my DBLP record and Google Scholar page.
Here below is a word-cloud (obtained from the titles of my research papers from IITB) describing my research interests:
- Image Restoration: This involves undoing the effect of natural degradation induced in images during or after acquisition. This includes tasks such as denoising, demosaicing, deblurring, inpainting, source separation.
- Tomographic reconstruction under constraints such as small number of measurements or low signal dosage or unknown/inaccurately known viewing parameters, exploiting different object-related constraints
- Compressed sensing: algorithms as well as theoretical error bounds, issues of erroneously specified sensing matrices, effect of different noise models, design of compressed sensing matrices
- Group testing/pooled testing algorithms
- Biomedical image reconstruction: CT, cryo-electron microscopy, diffusion MRI
- Low rank matrix completion, Robust PCA
- Image and video compression
- Sparse representations and dictionary learning
- Automated Turing Tests