An Innovative SIFT-Based Method for Rigid Video Object Recognition

Figure 10

The experimental results show that using the selected areas in bounding boxes in the left images as the training image, the other objects on the right are recognized correctly in corresponding video by RVO-SIFT eventually.