Two physical implementations of compressive sensing

Emmanuel Candes
Department of Mathematics and Department of Statistics
Stanford University


Wednesday, Apr 27, 2011
4:30 - 5:30 PM
Packard 101

Abstract:

Compressive sensing is a novel mathematical theory which asserts that one can recover signals or images of interest with far fewer sensors/measurements than were thought necessary. Since the theory may now be well established, one can imagine that a significant fraction of future research in the field will concern the implementation of this theory into -- hopefully -- effective hardware. This talk will report on such progress in two distinct areas. First, we discuss novel analog-to-digital converters which can operate without much information loss at a much lower sampling rate than that predicted by the Shannon/Nyquist theory. Second, we introduce an experimental setup for implementing compressive sensing ideas in biological imaging. Here, confocal microscopes provide high-resolution images but the imaging frame rate is typically slow and the sensibility is largely reduced compared to wide field imaging. Our experimental results demonstrate that it is possible to acquire certain types of high-quality microscopy images by taking far fewer measurements. Our aim is to discuss implementation challenges as well as the potential and limitations of compressive sensing technologies.

This is joint work with various scientists and engineers who shall be properly introduced during the talk.

---

Operations Research Colloquia: http://or.stanford.edu/oras_seminars.html