Journal of South China University of Technology (Natural Science Edition) ›› 2013, Vol. 41 ›› Issue (7): 107-113.doi: 10.3969/j.issn.1000-565X.2013.07.018

• Computer Science & Technology • Previous Articles     Next Articles

A Dynamic Time Warping Algorithm for Recognition of Multi- Stroke On- Line Handwritten Characters

Harold Mouchère1 Li Jin- peng1 Christian Viard- Gaudin1 Chen Zhao- xin1,2   

  1. 1. IRCCyN Laboratory,LUNAM Université,Université de Nantes,Nantes 44300,France; 2.School of Electronic andInformation Engineering,South China University of Technology,Guangzhou 510640,Guangdong,China
  • Received:2013-03-01 Online:2013-07-25 Published:2013-06-01
  • Contact: Harold Mouchère(born in 1981),male,ssociate professor,ainly researches on machine learning,n-line handwritingrecognition,ocument analysis and currently on on-line handwritten math expressions. E-mail:harold.mouchere@univ- nantes.fr
  • About author:Harold Mouchère(born in 1981),male,ssociate professor,ainly researches on machine learning,n-line handwritingrecognition,ocument analysis and currently on on-line handwritten math expressions.

PDF

282

Abstract:

Dynamic time warping (DTW) is a famous algorithm to compare two mono- stroke symbols online.It obeys boundary and continuity constraints and possesses high recognition rate.The extension of DTW to multi-stroke symbols raises specific problems.A naive solution is to convert the multi- stroke symbol into a single one via a direct concatenation in the handwriting order.However,people may write a symbol in different stroke di-rections and orders.The application of a brute force method by searching all the possible directions and orders leads to prohibitive calculation times.To reduce the searching complexity,DTW A * algorithm,which keeps the continuity constraint during each partial matching and reduces the searching complexity by using the A * al-gorithm,is proposed and is then tested on a flowchart dataset containing multi- stroke symbols.The results indi-cate that DTW A * algorithm helps to achieve the highest recognition rate and stability,as compared with such two other algorithms as DTW and modified Hausdorff distance.

Key words: pattern recognition, handwriting recognition, dynamic time warping