Figure 1: Tracking in an infrared (IR) image sequence. The contrast along the moving boundary is weak, particularly in the portion between the car and the street. There is also partial occlusion behind the tree toward the end of the sequence. Top three rows: tracking with the proposed kernel-based intensity tracking term. Fourth row: tracking in [13] with the Gaussian assumption. Bottom row: tracking by distributions with the method in [10]. Under each image is the frame number.