 |
 |
Students, Ph.D. and PostDoc candidates are welcome to contact me for further details about the projects or to present their own ideas for the projects they would like to do with me in computer vision and recognition.
Recognition
Object category recognition, scene classification, object detection and localization. Here is an illustration of a classification system and object class detection (face).
MoD Grand Challenge - visual recognition system for detecting threats by autonomous robots (Swarm). Best Innovative Idea for Swarm Team.
Visual Object Classes Challenge 2008 - visual recognition system. Best results by SurreyUvA_SRKDA Team.
TRECVid Challenge 2008 - visual concept retrieval system. Best results by the MediaMill Team.
Image representation
Detection of local features, image descriptors, invariance. Here is an illustration of scale and affine invariant feature detection and edge detection.
Image and video feature extraction
Invariant feature detection and local descriptors available at Featurespace.
Image retrieval
Indexing and retrieval of images and videos, image databases, video exploration and annotation. Here is an illustration of image retrieval and a demo applet.
Matching and registration
Finding correspondences between images.Here is an illustration of matching and a demo applet.
Clustering and data structures
Efficient clustering methods, nearest neighbor search, range search in high dimensional spaces.
Segmentation
Efficient segmentation color and grayvalue based methods for mid-level recognition cue extraction.
Image enhancement techniques
Sperresolutions, high dynamic range imaging, filtering, sampling and interpolation.
Video restoration, removal of blotch, scratch, flicker, unsteadiness.
