No abstract

Figure 1: The Light Chisel: (left) Prototype of our device. (left middle) User interaction with the Light Chisel in an augmented reality setup, (right middle) a close-up of our direct 3D modeling application work space, and (right) improvement of the 6DOF pose by template matching, showing the difference between the Light Chisel diffuse cylinder model and the physical tip projection in a camera image. AbstractWe present a novel interaction device tracked in 6 degrees of freedom by two commodity cameras. The inexpensive Light Chisel is statically illuminated with two LEDs, and uses no additional sensor (e.g. inertial or magnetic) or means of communication or synchronization. Its form factor is well suited for a screwdriver or chisel grip, allowing the Light Chisel to be rolled between the fingers. The position and orientation of the tool is tracked absolutely, making the Light Chisel suited for complex interaction, e.g. geometric modeling in augmented reality. The Light Chisel is physically small, limiting the physical and optical collisions with the real world. The orientation of the tool is tracked in a wide range of angles: pitch and yaw ±90 • , roll ±180 • . We evaluated our system against the OptiTrack optical tracking system. Our system achieved mean differences from OptiTrack reference of 2.07 mm in position, 1.06 • in yaw and pitch, and 5.26 • in roll using a pair of VGA cameras. We demonstrate usefulness of our Light Chisel in four applications: character animation, modeling by swirls, volumetric modeling, and docking of CAD models.

Abstract. Robust and accurate automatic detection of anatomical features on organic shapes is a challenging task. Despite a rough similarity, each shape is unique. To cope with this variety, we propose a novel clustering-based feature detection scheme. The scheme can be used as a standalone feature detection scheme or it can provide meaningful starting points for surface analyzing-based detection algorithms. The scheme includes the identification of a representative set of shapes and the usage of a specialized iterative closest point algorithm for the registration of shapes, which is followed by the projection of the features using the transformation matrix of the registration. Evaluation is based on a large set of expert annotated shapes and showed superior performance compared to state-of-the-art surface analyzing methods. Accuracy increased of 32 % and detection of all features is ensured.

Abstract. We propose a novel framework for the personalized design of organic shapes that are constrained to exhibit conformity with the underlying anatomy. Such constrained design is significant for several applications such as the design of implants and prosthetics, which need to be adapted to the anatomy of a patient. In such applications, vaguely defined work instructions are usually employed by expert designers to carry out a sequence of surface modification operations using interactive CAD tools. Our approach involves the abstraction of the work instructions and the expert knowledge into feature dependent machine interpretable rules in a Knowledge Base. Robustly identified canonical set of anatomical features are then employed to determine concrete surface shaping operations by a Smart Shape Modeler. These operations are eventually performed sequentially to adapt a surface to a target shape. The versatility of our approach lies in a priori defining an entire design workflow through a scripting language, thereby yielding a high degree of automation that is completely flexible and customizable via scriptable rules. Consequently, it eliminates tedious manual intervention and offers desirable precision and reproducibility. We validate this framework with a practical application -automatic modeling of shells in hearing aid (HA) manufacturing (HAM).