Please use this identifier to cite or link to this item: https://scholarbank.nus.edu.sg/handle/10635/111258
DC FieldValue
dc.titleIntegrated multiple neural network architecture for reading alphanumeric characters in complex scenes
dc.contributor.authorWan, Kong-Wah
dc.contributor.authorTian, Qi
dc.contributor.authorLow, Kah-Chan
dc.contributor.authorLau, Soo-Leng
dc.contributor.authorLui, Ho-Chung
dc.date.accessioned2014-11-27T09:46:18Z
dc.date.available2014-11-27T09:46:18Z
dc.date.issued1994
dc.identifier.citationWan, Kong-Wah,Tian, Qi,Low, Kah-Chan,Lau, Soo-Leng,Lui, Ho-Chung (1994). Integrated multiple neural network architecture for reading alphanumeric characters in complex scenes. IEEE International Conference on Neural Networks - Conference Proceedings 7 : 4384-4389. ScholarBank@NUS Repository.
dc.identifier.urihttp://scholarbank.nus.edu.sg/handle/10635/111258
dc.description.abstractAn integrated multiple neural network (NN) architecture is proposed to handle the automatic extraction and recognition of machine-printed alphanumeric characters in a complex scene. The computational framework is illustrated by our implementation of a multi-expert, multi-level reading system comprising of five NNs for both segmentation and recognition. A window-based alphanumeric NN (WNN_AD) is trained as a first-pass character-level primary recognizer. A window-based alphabet NN (WNN_A) and a window-based digit NN (WNN_D) are then used as group-level secondary recognizers. To further boost recognition accuracy, a Fourier-Descriptor-based Probabilistic neural network (FDPNN) is trained on characters which were misrecognized by the above three networks. In the event of mis-segmentation of merged or damaged characters, a special Character Horizontal Position Neural Network (CHPNN) is used to detect the horizontal positional occurrence of characters. Our system has been applied to the recognition and verification of the 11 machine-printed alphanumeric ID on containers. In spite of the variations due to character fonts, distortions, background color, image contrast, illuminations and ID layout formats, etc, the system has achieved a container ID-based recognition rate of more than 90% over 3000 images.
dc.sourceScopus
dc.typeConference Paper
dc.contributor.departmentINSTITUTE OF SYSTEMS SCIENCE
dc.description.sourcetitleIEEE International Conference on Neural Networks - Conference Proceedings
dc.description.volume7
dc.description.page4384-4389
dc.description.coden176
dc.identifier.isiutNOT_IN_WOS
Appears in Collections:Staff Publications

Show simple item record
Files in This Item:
There are no files associated with this item.

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.