Figure 1.
SVM scores for sequence- and frame-level ground truth. The upper pictorial representations are the video frames corresponding with the crosses on the respective SVM score plots below. Reprinted with permission from Ashraf et al., 2009.
SVM scores for sequence- and frame-level ground truth. The upper pictorial representations are the video frames corresponding with the crosses on the respective SVM score plots below. Reprinted with permission from Ashraf et al., 2009.