CaltechAUTHORS
  A Caltech Library Service

Multiscale Random Projections for Compressive Classification

Duarte, Marco F. and Davenport, Mark A. and Wakin, Michael B. and Laska, Jason N. and Takhar, Dharmpal and Kelly, Kevin F. and Baraniuk, Richard G. (2007) Multiscale Random Projections for Compressive Classification. In: 2007 IEEE International Conference on Image Processing. Vol.VI. IEEE , Piscataway, NJ, pp. 161-164. ISBN 978-1-4244-1436-9. https://resolver.caltech.edu/CaltechAUTHORS:20170424-164945669

[img] PDF - Published Version
See Usage Policy.

410kB

Use this Persistent URL to link to this item: https://resolver.caltech.edu/CaltechAUTHORS:20170424-164945669

Abstract

We propose a framework for exploiting dimension-reducing random projections in detection and classification problems. Our approach is based on the generalized likelihood ratio test; in the case of image classification, it exploits the fact that a set of images of a fixed scene under varying articulation parameters forms a low-dimensional, nonlinear manifold. Exploiting recent results showing that random projections stably embed a smooth manifold in a lower-dimensional space, we develop the multiscale smashed filter as a compressive analog of the familiar matched filter classifier. In a practical target classification problem using a single-pixel camera that directly acquires compressive image projections, we achieve high classification rates using many fewer measurements than the dimensionality of the images.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
http://dx.doi.org/10.1109/ICIP.2007.4379546DOIArticle
http://ieeexplore.ieee.org/document/4379546/PublisherArticle
ORCID:
AuthorORCID
Baraniuk, Richard G.0000-0002-0721-8999
Additional Information:© 2007 IEEE. Supported by NSF, ONR, AFOSR, DARPA and the Texas Instruments Leadership University Program. Thanks to Texas Instruments for providing the TI DMD developer’s kit and accessory light modulator package (ALP) and to Petros Boufounos for helpful discussions.
Funders:
Funding AgencyGrant Number
NSFUNSPECIFIED
Office of Naval Research (ONR)UNSPECIFIED
Air Force Office of Scientific Research (AFOSR)UNSPECIFIED
Defense Advanced Research Projects Agency (DARPA)UNSPECIFIED
Texas InstrumentsUNSPECIFIED
Subject Keywords:Data Compression, Image Coding, Image Classification, Object Recognition
DOI:10.1109/ICIP.2007.4379546
Record Number:CaltechAUTHORS:20170424-164945669
Persistent URL:https://resolver.caltech.edu/CaltechAUTHORS:20170424-164945669
Official Citation:M. F. Duarte et al., "Multiscale Random Projections for Compressive Classification," 2007 IEEE International Conference on Image Processing, San Antonio, TX, 2007, pp. VI - 161-VI - 164. doi: 10.1109/ICIP.2007.4379546
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:76882
Collection:CaltechAUTHORS
Deposited By: Kristin Buxton
Deposited On:25 Apr 2017 03:00
Last Modified:15 Nov 2021 17:03

Repository Staff Only: item control page