SURF and Texture Classification on Occluded Hand Images

Some experiments using SURF and texture classification on occluded hand images are presented in this paper. The region of interest (ROI) is usually involved in palmprint classification, but in our original experiments palm shape and ROI are missing. For each image we have used texture features, shape features and SURF descriptors to characterize it. Experiments are made on two datasets. The results obtained using SURF descriptors are better than those obtained using shape and texture, but it has the disadvantage of requiring long computing time. Fortunately, it is possible to reduce the computing time using a two steps method that combines shape or texture features and SURF descriptors, method described in this paper. This approach led to good results in less computing time.