Chakra Chennubhotla

Chakra Chennubhotla
Contact	Dept. of Computer Science University of Toronto Toronto, ON, CANADA M5S 3G4 chakra AT cs.toronto.edu
Research	COMPUTER VISION AND MACHINE LEARNING.
Focus	To identify, and represent, structure inherent to high-dimensional object-specific ensembles of visual, and other perceptual, data.
Recent Work	Hierarchical Eigensolver for Transition Matrices in Spectral Methods Chakra Chennubhotla and Allan Jepson. NIPS 2004 Spectral Embedding and Min-Cut for Image Segmentation [Poster] Paco Estrada, Allan Jepson and Chakra Chennubhotla. BMVC 2004 Perceptual Distance Normalization for Appearance Detection Chakra Chennubhotla and Allan Jepson. ICPR 2004 Hierarchical Representation of Transition Matrices for Spectral Clustering Chakra Chennubhotla and Allan Jepson. 2003 Algorithm/Demo/FAQ. EigenCuts: Half-Lives of EigenFlows for Spectral Clustering Chakra Chennubhotla and Allan Jepson. NIPS 2002 Sparse PCA (S-PCA): Extracting Multi-Scale Structure from Data Chakra Chennubhotla and Allan Jepson. ICCV 2001 Algorithm/Demo/FAQ. Facelets, shown below, are multi-scale orthogonal S-PCA basis derived from a database of face images. Unlike Eigenfaces, facelets represent fine-scale information with spatially local support. S-PCA basis derived from a database of gesturing hand images. Sparse Coding in Practice Chakra Chennubhotla and Allan Jepson. SCTV 2001 Robust Contrast-Invariant EigenDetection Chakra Chennubhotla, Allan Jepson, and John Midgley. ICPR 2002 Spectral Clustering of Protein Sequences Alberto Paccanaro, Chakra Chennubhotla, James Casbon and Mansoor Saqi. IJCNN 2003 Markov Analysis of Protein Sequence Similarities Chakra Chennubhotla and Alberto Paccanaro. 2003