Graduate Student / Postdoc Seminar

A Symbol-Based Algorithm for Decoding Noisy Bar Codes

Speaker: Rachel Ward, Courant

Location: Warren Weaver Hall 1302

Date: Friday, April 15, 2011, 1 p.m.


Bar code reconstruction involves recovering a clean signal from an observed signal that is blurry and corrupted by additive noise. The precise form of the blur kernel is unknown, making reconstruction harder than standard deblurring. On the other hand, the set of valid bar codes is very small relative to the set of all binary sequences, and this additional information makes reconstruction feasible. In this talk we show how bar code reconstruction can be re-cast as a sparse-recovery problem, and we develop a fast symbology-based reconstruction algorithm. This is joint work with Fadil Santosa and Mark Iwen.