Font Size: a A A

Adaptive Expand Of Chinese Font Based On ARM7 Embedded System

Posted on:2009-05-14Degree:MasterType:Thesis
Country:ChinaCandidate:Z N HanFull Text:PDF
GTID:2178360272479531Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
The storage space and addressability of embedded system is very valuable. It would be a restriction when the Chinese fonts storage technology and Chinese character recognition in computer system were used in embedded system. So it would be very important to study with the Chinese fonts storage technology and Chinese character recognition for embedded system. The importance of the Chinese font in the Chinese character recognize technology based on the embedded system is deeply discussed. The method of Chinese fonts design is summarized based on the theory of the Chinese font movement in Chinese font technology. Above all a new method to recognize all the Chinese in one font with fewer Chinese fonts is proposed. It is a good way to reduce the waste of resource for the embedded system.The combination of Wavelet transform and artificial neural network is a new method in the Chinese character recognize field such as the recognition of numeral character, letter character, font, vehicle plate, but this method has localization because of the character of the printed chinese font. The directional vector is added in the ridgelet transform based on the normal wavelet transform. Extracted the strokes vector of the Chinese character picture in the direction of Heng,Shu,Pie,Na, and put these vector into BP neural network to recognize the Chinese character in the font of "Simsun". This is the innovation of the application in ridgelet transform and the algorithm of the Chinese character recognition. The artificial neural network has the ability to approach arbitrary non-linear function and the features of self-study and self-organization. Use the less number of trained Chinese font of "Simsun" to recognize all the Chinese font of "Simsun" is achieved. The resource of the embedded system is saved. And the row and queue weighted encoding method can now be combined with to recognize Chinese character in the embedded system. Experiment shows that this method can recognize the Chinese character effectively.
Keywords/Search Tags:character recognition, ridgelet transform, BP neural network, Chinese strokes feature, row and queue weighted encoding
PDF Full Text Request
Related items