Sketch recognition based on manifold learning
Heeyoul Choi, Tracy Hammond · 2008
Current feature-based methods for sketch recognition systems rely on human-selected features. Certain ma-chine learning techniques have been found to be good nonlinear features extractors. In this paper, we ap-ply a manifold learning method, kernel Isomap, with a new algorithm for multi-stroke sketch recognition, which significantly outperforms the standard feature-based techniques.