Non-Convex Shredded Signal Reconstruction via Sparsity Enhancement
Non-Convex Shredded Signal Reconstruction via Sparsity Enhancement

Submitted in: 42nd IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), March 2017
Paper Name: Non-Convex Shredded Signal Reconstruction via Sparsity Enhancement
Authors: Arindam Bose, Mojtaba Soltanalian
Abstract: Restoration of shredded signals remains a relevant and significant challenge in archaeological and forensic efforts. In this work, we present a novel approach for reconstruction of shredded signals (including text documents and images) within a context of general multidimensional sparse signals. To this end, we present a generic efficient non-convex optimization method that employs iterative sparsity enhancement of the observed signal. A key component of the design follows from the observation that most natural signals are sparse in a given representation domain. Computational results portrait the potential of our suggested method in several practical cases of signal reconstruction.
Keywords: Non-convex optimization, signal reconstruction, sparsity, strip-shredded documents.
Date of Conference: 5-9 Mar. 2017.
Conference Location: New Orleans, LA, USA.

About the author

Scientific History Blog Writer • Art enthusiast and Illustrator • Amateur Photographer • Biker and Hiker • Beer Enthusiast • Electrical Engineer • Chicago

Leave a Reply

Twitter Feed
%d bloggers like this: