| Summary: | The Chinese character recognition (Chinese OCR) aims at freeing people from inefficient and tedious data input, creating values in document analysis. This technique has a wide application in the real world, such as postal address recognition, signature verification, industrial parts inspection, data mining and office automation. Now a growing need for such technique has been seen, as Chinese communities are already deep in the information era. The application feasibility has been much enhanced because of the accumulation of productive researches in recent years and cost reduction driven down by the technological advances of computer industry.
In this thesis, A new structural representation for Chinese characters, and preclassification and fuzzy matching recognition scheme based on this representation are proposed for off-line Chinese characters. Like most of the recognition system, the proposed recognition system consists of four basic parts: pre-processing, feature extraction, pre-classification and recognition.
|
|---|