Actually, this is object detection because there is a bounding box which is generated localizing the object(s). Classification would be if each frame returned

Actually, this is object detection because there is a bounding box which is generated localizing the object(s). Classification would be if each frame returned a class and class score without the bounding box.