Wednesday, October 29, 2008

CS: Comment on non uniform sampling, Digital Rain, a post-doc, a thesis

Most of today's entries were found on the interwebs:

....(2) According to Holger Rauhut: "The ideas of compressed sensing are not completely new. There are actually many precursors. For instance, the use of [l.sub.1] minimization seems to be first mentioned in the Ph.D. thesis of Logan in 1965. [l.sub.1] minimization was further used by geophysicists starting in 1970s. In statistics the field of variable selection has introduced [l.sub.1] minimization (called LASSO) and greedy algorithms (stepwise, forward regression, and projection pursuit regression, ere) in the 1980s and 1990s. The new contribution of compressed sensing consists in the type of applications, and in rigorous (and often fascinating) proofs of recovery results using new concepts and connections to other fields of mathematics." The paper by [17] is, in a sense, compressed sensing for nonlinear transformations....

...[17] F. Marvasti and A. K. Jain, Zero crossings, bandwidth compression, and restoration of nonlinearly distorted band-limited signals, Journal of Opt. Soc. of America, 3(5), 651-654, 1986....
There is a new field where CS is being used as featured in One Video Stream to Serve Diverse Receivers by Szymon Chachulski, Dina Katabi and Grace Woo. The abstract reads:
The fundamental problem of wireless video multicast is to scalably serve multiple receivers which may have very different channel characteristics. Ideally, one would like to broadcast a single stream that allows each receiver to benefit from all correctly received bits to improve its video quality. We introduce Digital Rain, a new approach to wireless video multicast that adapts to channel characteristics without any need for receiver feedback or variable codec rates. Users that capture more packets or have fewer bit errors naturally see higher video quality. Digital Rain departs from current approaches in two ways: 1) It allows a receiver to exploit video packets that may contain bit errors; 2) It builds on the theory of compressed sensing to develop robust video encoding and decoding algorithms that degrade smoothly with bit errors and packet loss. Implementation results from an indoor wireless testbed show that Digital Rain significantly improves the received video quality and the number of supported receivers.

A PostDoc has just been listed on the CSjob section. It is a FP6/Marie Curie Fellowship for a Post-doctoral fellow in Sensor Networks & Compressive Sensing at the Institute of Computer Science (ICS), Foundation for Research and Technology-Hellas (FORTH) in Heraklion, Crete, Greece. The position focuses on information theoretical aspects of sensor networks and Bayesian compressive sensing. The ideal candidate is expected to have expertise in information theory, signal processing, compressive sensing, pattern recognition and data fusion, statistical communications, and preferably experience in engineering applications concerning the structure, function, and organisation of WSN systems. Job starting date: 01/01/2009. Aplication deadline: 01/12/2008.

I also found this, but did not find the actual document. Análisis y reconocimiento de patrones en electroforesis capilar utilizando compressed sensing by Alvaro Hernández Orence, Hernández Orence, Alvaro, Paredes Quintero, José Luis at Universidad de Los Andes in Venezuela. CS is electrophoresis, I'd like to see more of that.

There is a call for a Proceedings of the IEEE Special Issue on: Applications of Sparse Representation and Compressive Sensing. But it should be noted that:


JackD said...

To complete Holger Rauhut's comment on the greedy approach, Matching Pursuit, one of the first greedy method seen as way to reach the sparse representation of a signal in a given dictionary, goes back to 1938 ! (see here).

In fact, I'm wondering if Gauss, or even Euclid, hadn't designed an iterative method very close to MP (solving "Ax = b") ;-)

Igor said...

Thanks Laurent.