Raising hand is one of the most important types of interaction between students and lecturers in classroom. When an automatic system can be installed in classroom to figure out which students raise their hands, it is possible to design more advanced applications for education goals. This paper proposes a system that employs computer vision technologies to automatically detect the student action of raising hand. We first design a foreground extraction method to segment student bodies in consecutive video frames. Next, a shape-like appearance signature that represents human gestures is designed based on the scale-invariant feature transform (SIFT) descriptor. A gesture classifier for raising hands is also designed using the support vector machine (SVM) approach. This paper designs several experiments to demonstrate the performance of our proposed system in a real classroom.