Handwritten Character Recognition Using Nonsymmetrical Perceptual Zoning

Abstract

In this paper we present an alternative strategy to define zoning for handwriting recognition, which is based on nonsymmetrical perceptual zoning. The idea is to extract some knowledge from the confusion matrices in order to make the zoning process less empirical. The feature set considered in this work is based on concavities/convexities deficiencies, which are obtained by labeling the background pixels of the input image. To better assess the nonsymmetrical zoning we carried out experiments using four different zonings strategies. Experiments show that the nonsymmetrical zoning could be considered as a tool to build more reliable handwriting recognition systems.

Tree Architecture Pattern Distributor: a task decomposition classification approach

“…So, it was developed a new automatic method which extracts information from the confusion matrices. Freitas et al [12] showed that it is possible to arrange the similar classes based on the confusion caused by a classifier. Table I shows a hypothetic confusion matrix of a 8-class problem.…”

“…Freitas et al [12] defined manually the class grouping using the confusion matrix (CM). In order to have the first CM, they trained an ordinary single neural network.…”

Text and Non-text Segmentation and Classification from Document Images

“…Zoning, usually being applied tro alphanumeric character recognition [28], [29], [30] and [31], is a technique where an n x m grid is superimposed on the binary image and information for each n x m grid is computed. This research focuses on the value of the average of the white pixels that represent the foreground of the image in each zone.…”

Distance-based Disagreement Classifiers Combination

“…Based on this and other works we have done [3]- [6] we have chosen this architecture for our experiments. Moreover, experimental results support this architecture.…”

“…The feature set is obtained by labeling the background pixels of the input image as belonging to either a concavity or a convexity region [12], as presented in Figure 2b. The alphabet of symbols was adapted to handwritten characters, resulting in 24 different symbols [3]. Several authors have presented zoning mechanisms or regional decomposition methods to investigate the recognition of patterns from their parts, similarly to what the human brain does during the reading process.…”