1 November 1997 Off-line writer verification utilizing multiple neural networks
Author Affiliations +
Optical Engineering, 36(11), (1997). doi:10.1117/1.601550
Abstract
A writer verification system based on multiple neural network classifiers is described, aimed to be easily extendable. Each writer registers with the system by writing a set of discrete Chinese characters. One neural network is constructed for each enrolled character class on a per-person or a per-subgroup basis. The decisions from individual network classifiers are combined by voting. The method has been verified to work reliably on our testing database: a verification rate above 96% is achieved on short-term data.
Kai Huang, Jing Wu, Hong Yan, "Off-line writer verification utilizing multiple neural networks," Optical Engineering 36(11), (1 November 1997). http://dx.doi.org/10.1117/1.601550
JOURNAL ARTICLE
7 PAGES


SHARE
KEYWORDS
Neural networks

Feature extraction

Databases

Optical engineering

Image segmentation

Binary data

Distance measurement

RELATED CONTENT

Reduction Of The Scattering Loss Of Polycrystalline Fibers
Proceedings of SPIE (October 16 1984)
Dynamic response in carbon fiber composite material
Proceedings of SPIE (May 31 1999)
Optical multilayer neural networks
Proceedings of SPIE (November 01 1991)
Effect of experimental design on sample size
Proceedings of SPIE (July 01 1991)

Back to Top