On-line Sketching Recognition For Conceptual Design

On-line recognition is a branch of pattern recognition and computer vision, and plays a very important role in Human-Machine Interactions. It studies the methods and techniques to recognize a sketch as it is drawn using a digitizer or an instrumented stylus.With the rapid growth of the computer technology and its widespread acceptance to the society, this research draws much more attention recently, many research teams devote to this complex task which covers a broad field. The present research results mainly concentrate in the hand printed characters recognition, but the freehand sketching recognition research is just started. Aiming at the defects of the existing methods, a study on developing an effective recognition method is performed in this dissertation.In this thesis, we mainly study the method for on-line recognizing the freehand sketching for the single stroke. The stroke is classified and defined based on the geometry characteristic. We adopt the adaptive polygonal approximation approach and least median square curvilinear regression to recognize the single linear element. The recognition results include line, fold-line, ellipse, elliptic arc, circle, circular arc, hyperbola and parabola. Based on the characteristics of the segment points, classify the compound linear element as the ultra compound linear element and the non-ultra compound linear element, and we conclude the segmentation algorithms accordingly. The freehand sketch recognition prototype system is independently developed in the above method and confirms our algorithms to be serviceable and extensible.This dissertation is the project supported by Natural Science Foundation of Jiangxi Province (No.0311018) and Open Foundation of Research Center of Measuring Technologies and Control Engineering of Jiangxi Province (No.2003-013).