Multiscale Random Projections for Compressive Classification
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.
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.Attached Files
Published - 04379546.pdf
Files
Name | Size | Download all |
---|---|---|
md5:fc663f7d157579f10fb68ba9a93c54f8
|
410.9 kB | Preview Download |
Additional details
- Eprint ID
- 76882
- Resolver ID
- CaltechAUTHORS:20170424-164945669
- NSF
- Office of Naval Research (ONR)
- Air Force Office of Scientific Research (AFOSR)
- Defense Advanced Research Projects Agency (DARPA)
- Texas Instruments
- Created
-
2017-04-25Created from EPrint's datestamp field
- Updated
-
2021-11-15Created from EPrint's last_modified field