Automatic Teaching System Of Ligaturing-letters Based On The Writing Process Tracking

English writing is a major focus of teaching mission in primary school. To meetthe daily use of English writing, Be able to help students writing English neatly,standardized and smoothly has become a qualified claim in primary schools which hasthe conditions. To solve the problem of shortage of educational resource, to relievethe stress of teachers, to response the call from our nation for the establishment ofenvironment-friendly society, and to keep pace with the development of theinformation society, The develop of automatic teaching system—includingligaturing-letters teaching system—has become a hot research area. We usually writeligaturing-letters when we write western languages such as English, even if thestructure of ligaturing-letters no that simple, it still have its own rules. Usingcomputer to achieve the automatic teaching algorithm has these following difficulties:1. The more letters are ligatured, the more writing difficulty is increased.2. The morecomplex the monomer letter is, the more complex the ligaturing-letters will be.3. Thelonger cursive stroke will increase the number of extracted strokes, and makes themdiversification. How to use computer to achieve the automatic teaching algorithm isthe seriously problem in our project. For the difficulty, series of experiments havebeing done and those achievements and innovations were obtained:First, for different writing ways of English letter-ligaturing, divide ligaturing-letters type into direct oblique ligatures, direct cross ligatures, close oblique ligaturesand no-linking ligatures totally four categories. Meanwhile, analyze the actual Englishwriting of students to exact control of their writing features and error-prone points.Due to the above, we propose the concepts of synchronous tracking, analysis andguidance of guided English cursive writing system based on touchscreen.Second, propose the synchronous extract method of handwriting features. Afteranalyzing more than200kinds of ligaturing-letters strokes, A new stroke trackingalgorithm which based on the key points of stroke segmentation was proposed, Eachkinds of strokes’ key points was acquired from figure domain and time domain, andwere used to extract the stroke. The method of extracting key points from doubledomain ensure the extract of characteristic and number of those key points.Third, using curve fitting method. To divide template strokes and sample strokes,each kinds of template strokes’ key points was used to extract the stroke and acquire condemn of template strokes could used to curve fitting. Conic curve, sinusoid andline has been chosen as standard fitting curve. And the best fitting-curve’s coefficientvector template and the coefficient of determination will obtained by least squares.The key points form sample writing and fitting curve(reconstructed by coefficientvector) will used to compute the similarity. At last, the system will output the writingguidance or final score.At last, for the sensitive feature of the spatial logic caused by less number ofstrokes which unique in English ligatures, a adaptable strokes relations division andanalysis process has been proposed. We put forward a targeted description of thealgorithm for simple spatial relationships and intersection relationships betweenstrokes. Meanwhile, the system will set the appropriate error message where thelogical errors of spatial relationships may exist on.The experiment shows that the method is objective and robust, besides, it can beapplied to more than two Ligaturing-Letters Stroke Tracking and other kinds ofwestern language.