Automatic landmarking has overcome a main drawback in Active Appearance Models (AAMs) computer vision technique: landmarks must be manually placed on training images to construct the shape mesh. Although AAMs are a well-known procedure for statistical matching of object shape and texture between images, hand-landmarking makes it very time consuming and not automatically applicable on new objects observed in images. There is a vast body of work applying automatic landmarking on faces or body joints for AAM training and several other purposes. In this paper, first we explore the possibility to extend one of these methods to full body contours on still images supplied by a single camera and demonstrate it is a plausible approach in terms of accuracy and speed measures from experimentation. Then, a 3D upgrade approach is presented using a RGB-D sensor to detect and automatically landmark body shapes with high accuracy in real-time, as a comparison with the latest technology. Our proposal represents a new research line in human body pose tracking with a single-view camera and the first steps of a novel contribution in learning body appearance. Hence, further implementation in robots would lead to people being recognized and identified by them with any vision resources, no matter how primitive or Advanced in human-robot interaction tasks.