Recognition Based On Neural Network Online Handwritten Pitman Shorthand

High-speed Data Entry is becoming more challenging in the recently years with the great progress in Electronic Computer and Worldwide web,especially in the area of Human Computer Interface. Many research topics had focused on how human can input information into a computer as quickly as possible with high accuracy rate in an easy way. The recognition of handwritten word is one possible solution. But the recognition of nature word writings encountered lots of problems,such as,low writing speed and low recognition rate.Shorthand is a high-speed recording technique. It has the advantage of high recording speed and brief strokes that can be easily recognized by certain artificial intelligence way. After nearly 200 years development,Pitman's shorthand is now a very robust shorthand recording system,and it's the most popular shorthand in the English world.This paper presents a new technique for segmentation and recognition of handwritten Pitman shorthand vocalized outlines. Due to its low redundancy,recognition of the Pitman Shorthand requires higher performance of outline segmentation and strokes classification. The proposed method in this paper raised a way of relatively high performance recognition of pitman shorthand vocalized outline. The approach includes:a) Segmentation of the vocalized outlines,including correction of the over segmentation based on neural networks.b) Feature extraction and recognition of Pitman shorthand consonant signs using neural networks.c) The result of a small test containing 68 most frequently used English words is given. The average accuracy rate of the test words reaches 89.6%.