Reconstruction of compressively sensed images via neurally plausible local competitive algorithms

Abstract—We develop neurally plausible local competitive algorithms (LCAs) for reconstructing compressively sensed images. Reconstruction requires solving a sparse approximation problem. Our solution technique uses a neural network that emulates how the brain may actually solve such sparse approximation problems. This method could ultimately be implemented in analog electronics, which would not only significantly diminish processing time, but also enable analog implementations for both acquisition and reconstruction. I. BACKGROUND A. Motivation A new technique called compressive sensing permits a signal to be captured directly in a compressed form rather than recording raw samples in the classical sense. With compressive sensing, only about 5-10% of the original number of measurements need to be made from the original analog image to retain a reasonable quality image. The theory of recording an image using compressive sensing is well established [3], [4], but much research remains i...

• Type: Conference
• Year: 2008
• Where: CISS
• Authors: Robert L. Ortman, Christopher J. Rozell, Don H. Johnson
• Keywords: CISS 2008, Compressive Sensing, Information Management, Local Competitive Algorithm, Sparse Approximation Problem