Published July 9, 1991
| Published
Book Section - Chapter
Open
Continuous-time segmentation networks
- Creators
- Harris, John G.
- Others:
- Mathur, Bimal P.
- Koch, Christof
Abstract
Segmentation is a basic problem in computer vision. The tiny-tanh network, a continuous-time network that segments scenes based upon intensity, motion, or depth is introduced. The tiny- tanh algorithm maps naturally to analog circuitry since it was inspired by previous experiments with analog VLSI segmentation hardware. A convex Lyapunov energy is utilized so that the system does not get stuck in local minima. No annealing algorithms of any kind are necessary- -a sharp contrast to previous software/hardware solutions of this problem.
Additional Information
© 1991 Society of Photo-Optical Instrumentation Engineers (SPIE). This work was done in close cooperation with Christof Koch. Thanks to Carver Mead for making this research possible. All chips were fabricated through MOSIS with DARPA's support. This research was partially supported by NSF grant IST-8700064, a grant from the Office of Naval Research, by DDF-II funds from the Jet Propulsion Laboratory at the California Institute of Technology.Attached Files
Published - 161.pdf
Files
161.pdf
Files
(844.9 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:50f8f33ffb63a3b7b89119e6ed6d9744
|
844.9 kB | Preview Download |
Additional details
- Eprint ID
- 91816
- Resolver ID
- CaltechAUTHORS:20181213-143633096
- Defense Advanced Research Projects Agency (DARPA)
- IST-8700064
- NSF
- Office of Naval Research (ONR)
- JPL/Caltech
- Created
-
2018-12-18Created from EPrint's datestamp field
- Updated
-
2021-11-16Created from EPrint's last_modified field
- Series Name
- Proceedings of SPIE
- Series Volume or Issue Number
- 1473