My current research focuses on the development of computational geometry algorithms for sparse coding. Recent results include the Greedy Basis Pursuit algorithm for computing signal representations, a geometric interpretation of sparse coding, and the successful application of clustering to dictionary learning.

I am currently developing related methods for compression, nonnegative matrix factorization, and linear programming. Stay tuned.

My earlier research centered on computer vision, specifically early vision. Topics included occlusion identification, edge detection, and shape-from-shading.