Imaging, Vision and Pattern Recognition (IVPR)

Image Processing, Computer Vision and Pattern Recognition constitute a fascinating area of research. In IVPR group at ETCE-JU, we are interested in both theoretical and practical aspects of the above fields.

As part of theoretical developments, we are exploring Graph-based modeling (graph cuts, graph matching, detection of shortest paths, optimum path forest, Spectral graph theory and Delaunay graphs), various statistical techniques (Particle filtering, non-parametric estimations and mixture models), PDEs (level sets) and techniques from signal theory (Fourier Transform, Hilbert Transform, Steerable maps).

On the application side, we are working on Multimedia analysis (video summarization and movie scene detection), Target tracking (pedestrian tracking, tracking of human monocyte cells and muscle stem cells), Medical image segmentation (segmentation of kidney, retinal blood vessels, cerebral white matter and blood vessels near the pancreas in zebra-fish), Medical image registration (registration of retinal images and brain MRI images), Computer aided detection (detection of pelvic fractures), Bio-metric authentication (subject invariant gait recognition from silhouettes) and Image Fusion (across different spectra).