In scenarios with visually similar targets, trackers that rely primarily on IoU and motion prediction (such as ByteTrack) are more suitable than trackers that heavily depend on appearance features.