Sketch recognition is one of the hinges of sketch understanding, and it is the foundation of sketch semantic extraction. Now, many institutions are working at sketch recognition. And there have been many sketch recognition approaches; however, the recognition accuracy and the inputting freedom still need to be improved.Support Vector Machine (SVM) is currently a popular machine learning algorithm based on the Statistical Learning Theory (SLT). SVM can efficiently solve small-sample machine learning problems. Now, there are few works of sketch recognition based on SVM, and the existing works have not take full advantage of SVM's powerful capability of learning, classification and generalization.This thesis introduces the actuality of sketch understanding, the algorithms and applications of SVM. A sketch recognition system based on SVM is implemented.The main works can be summarized as follows:First, we discuss the principle of SVM method, and we analyzing current sketch recognition methods.Second, we discuss the existing methods of sketch recognition and feature extraction, and we present the application of SVM in sketch recognition, and a prototype system is conducted.Finally, the application of the program flowchart demonstrates that SVM approach is feasible. |