A real-time neural system for color constancy
- Creators
- Moore, Andrew
- Allman, John
- Goodman, Rodney M.
Abstract
A neural network approach to the problem of color constancy is presented. Various algorithms based on Land's retinex theory are discussed with respect to neurobiological parallels, computational efficiency, and suitability for VLSI implementation. The efficiency of one algorithm is improved by the application of resistive grids and is tested in computer simulations; the simulations make clear the strengths and weaknesses of the algorithm. A novel extension to the algorithm is developed to address its weaknesses. An electronic system that is based on the original algorithm and that operates at video rates was built using subthreshold analog CMOS VLSI resistive grids. The system displays color constancy abilities and qualitatively mimics aspects of human color perception.
Additional Information
© 1991 IEEE. Manuscript received August 13, 1990; revised November 21, 1990. This work was supported (via fellowships to A. Moore) by the Parsons Foundation and the Pew Charitable Trust and (via research assistantships) by the Office of Naval Research, the Joint Tactical Fusion Program, and the Center for Research in Parallel Computation. The authors are grateful to many of our colleagues at Caltech and elsewhere for discussions and support in this endeavor: G. Fox, F. Perez, and S. Shein for discussions about color constancy; M. Mahowald, C. Mead, and M. Sivilotti, inventors of the original silicon retina, for systems and VLSI discussions; J. Hams, J. Luo, and C. Koch for discussions about resistive grids; D. Lyon, M. Mahowald, and S. Ryckebush for discussions about sample-and-hold circuitry; J. Lazzaro for discussions on systems issues; and s. Chascsa, T. Horiuchi, and F. Perez for assistance with photography. The authors express their gratitude to DARPA for MOSIS fabrication services, and to Hewlett Packard for computing support in the Mead Lab.Attached Files
Published - MooreAllmanGoodman1991.pdf
Files
| Name | Size | Download all |
|---|---|---|
|
md5:9edf4cbd201e9dcfac80a051956cec1c
|
1.3 MB | Preview Download |
Additional details
- Eprint ID
- 76343
- Resolver ID
- CaltechAUTHORS:20170408-172537077
- Ralph M. Parsons Foundation
- Pew Charitable Trust
- Office of Naval Research (ONR)
- Joint Tactical Fusion Office
- Center for Research in Parallel Computation
- Created
-
2018-03-13Created from EPrint's datestamp field
- Updated
-
2021-11-15Created from EPrint's last_modified field