An Active Shape Model Based Tactile Hand Shape Recognition with Support Vector Machines

This paper presents a novel approach to hand shape recognition problem with support vector machines (SVMs) by establishing a new ASM (active shape models) based kernel from the shape contours. This kernel takes advantage of ASM to model deformable shape contours and thus is more robust to noise and shape variations. By incorporating the similarity criterion employed in ASM, we introduce an ASM based kernel used for SVM classification, which in turn allow for considerable variability and have a more reasonable distance measure. The proposed method combined the strength of ASM shape searching and SVM discriminating and therefore achieve a better recognition rate than conventional template matching method.