A Normalization Research Of Converting Handwriting Character To Stroke-Segment-Mesh Glyph

Chinese character reorganization is an important sub-field of patternrecognition, which is also a most difficult issue to be fixed, involve thepattern recognition, image processing, digital signal processing, naturallanguage understanding, artificial intelligence, vague mathematics,informationism, computer science, Chinese language informationprocessing and so on. Chinese Character recognition is a integratedtechnology and is based on the preprocessing. The resultsof preprocessing directly impact on the effects of Chinese characterrecognition, so preprocessing plays a very important role in the study ofChinese character recognition.stroke-segment-mesh, which uses a line-segment of pre-defined lengthand direction as a glyph description element (stroke segment). Since it isequipped with suitable granular degree, free of ambiguity, and standardized, itcan describe the glyph skeleton of all Chinese characters (including wronglywritten characters, variant forms of characters in ancient literatures, andcombined-characters). It’s important that it can be applied in the computationfor comparing either the whole or part of glyphs. There are twoways to provide stroke-segment-mesh glyph. One is the descriptive input ofglyph in the stroke-segment-mesh; the other is handwriting character. Theformer way can get the stroke-segment-mesh glyph directly. However,because of line-segment of pre-defined length and direction, this way isunnatural, inconvenient and inefficient. The latter requiresa regularization process that machine is good at doing to get thestroke-segment-mesh glyph, but the handwriting’s writing is in line withpeople writing habits, natural, high efficiency, so that people and machine cando the things they are good at doing. The paper focuses on the preprocessing of handwritten character, therelative positional relationship among strokes, handwriting strokesnormalization processing and visual debugging platform. Specific activities asfollows:1. A detailed development and current situation and problems ofhandwriting recognition technology.2. Details of the preprocessing of on-line handwritten character, whichlaid a good foundation for extraction and regularization of the key points ofthe strokes3. Through the introduction of graphical relative positional relationship,the relative positional relationship among components of Chinese characterand the relative positional relationship among the strokes of Chinesecharacter, a fit relative positional relationship among the strokes of Chinesecharacter is proposed for the paper, the relationship directly effectthe use of normalization methods.4. Proposed two normalization methods. One method is that relationshipnormalization based on the strokes; the second is that strokes normalizationbased on the relationship. And established visual debugging platform that isgood for debugging program and improve the efficiency of programdevelopment.