Neural networks have been used to recognize handwritten characters such as Chinese, English or numerals. But their performance, i.e., the recognition rate, depends on a number of factors which may include the network architecture, feature selection, network parameter setting, learning strategy, learning sample selection, test pattern preprocessing, etc. These factors are important to network engineer in designing a network for a particular application problem, but unfortunately there is a lack of systematic way to guide their decision making regarding the selection of these parameters. This paper presents a parameter tuning (namely the selectivity parameter) methodology based on a sensitivity analysis of the Neocognitron model, and the off-line handwritten numeral recognition with supervised learning is chosen to be the demonstrated application problem. Genetic algorithm (GA) is used to select parameters leading to improved recognition results. We used a set of training pattern provided by Fukushima[S] as our training patterns which involved no preprocessing, and our experimental results show a significant improvement in performance. A brief discussion on alternate hybrid architecture involving neural network and genetic algorithm, and different fitting functions for the GA will be presented.
Off-line recognition of handwritten Chinese characters is a formidable challenge. Both structural and nonstructural approaches have only achieved limited success. While the non-structural approaches deal with the noise and minor structure-distortions by gathering a large number of character samples, structural ones perform better in handling large structural variations. Recently, neural network is also employed to offer alternative solutions in various manners. However, in these attempts neural network is seldom used to address the handling of the strong structural chmcteristics of a character set. This paper aims to incorporate structural knowledge of the character set into a neural network, to take advantage of both approaches. Two sets of tests have been carried out. One of them deals with the recognition of seventeen categories of simple Chinese characters. With ten samples in each category written by five individuals, the system achieves a recognition rate of 52.4%. Another test involves the expansion of the system to handle fifty categories. One sample set is used to demonstrate the expandability of this approach. In this test, the system still achieves a recognition rate of 62%. By this approach, potential in learning new writing styles and tolerance to input inexactness are readily achievable by the means of neural network settings. Moreover, techniques in handling large structure variations (as established i n structural approaches) can also be realized in the system.
In employing a structural approach to recognize handwritten Chinese characters (HCC), substroke is usually chosen as a feature set. Various types of ambiguities may arise during the substroke extraction, for instance, touchings among the substrokes or their distortions. Most conventional approaches would retain the most likely choice based on the "law of good continuity", while others may output all possible alternatives without offering any ranking information on the output candidate sets. A fuzzy substroke extractor was previously proposed by the authors to handle a number of ambiguities caused by substroke touchings, which measured the goodness of the "continuity law" by assigning weights to the extracted substrokes according to certain criteria. All detectable fuzzy substrokes were also retained. The extractor then made use of the scoring information on the extracted items to produce a set of "consistent" outputs, from which no edgesharing is found among the extracted items. The major advantage gained from this approach is that it is possible to reconstruct the skeleton in a "justfit" mode. Thus, for the candidate set with most of the desirable extracted substrokes, the noise rate is much lower than those reported by using a conventional "explore-all-possibility'' approach. To achieve this goal, the segmentation of character skeleton is transformed into a fuzzy set partitioning task. This paper extends our current technique to handle another type of ambiguity problem in substroke extraction, i.e. broken substrokes. Two cases of broken substrokes are addressed. Under certain conditions, virtual edges are introduced to "complete" the "supposed" broken skeleton graph. By considering this new skeleton as the actual input, suspected broken substrokes are detectable as well. With this proposed extension, most ambiguities encountered during substroke extraction can now be successfully treated in a unified framework.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.