Abstract: A system and method for recognizing instances of classes in a 2D image using 3D class models and for recognizing instances of objects in a 2D image using 3D class models. The invention provides a system and method for constructing a database of 3D class models comprising a collection of class parts, where each class part includes part appearance and part geometry. The invention also provides a system and method for matching portions of a 2D image to a 3D class model. The method comprises identifying image features in the 2D image; computing an aligning transformation between the class model and the image; and comparing, under the aligning transformation, class parts of the class model with the image features. The comparison uses both the part appearance and the part geometry.

Abstract: A robotic hand comprises a base having fingers and a palm. The fingers define a working volume. The hand is able to grasp objects between the fingers and the palm by varying a distance from the palm to the working volume. Varying this distance can be achieved by either moving the palm towards the working volume or by moving the working volume towards the palm. Robotic hands that move the palm comprise an actuator that extends the palm from the base towards the working volume. Robotic hands that move the working volume relative to the palm have the fingers mounted to a common finger support unit that is configured to translate relative to the base.

Abstract: A system and method for automatically computing desirable palm grasp configurations of an object by a robotic hand is disclosed. A description of the object's surface is obtained. A grasp configuration of a robotic hand comprising a palm and one or more fingers is selected, which specifies a hand configuration and joint variables of the hand, and a plurality of contact points on the object's surface for the palm and fingers. The coefficient of friction at each of the contact points is determined, and a description of one or more external wrenches to which may apply to the object is acquired. The ability of the robotic hand to hold the object against the external wrenches in the selected configuration is then computed. In some embodiments, a plurality of grasp configurations may be compared to determine which are the most desirable. In other embodiments, the space of hand configurations, or the smaller space of palm contact configurations, may be searched to find the most desirable grasp configurations.

Abstract: A system and method for performing object and class recognition that allows for wide changes of viewpoint and distance of objects is disclosed. The invention provides for choosing pose-invariant interest points of a three-dimensional (3D) image, and for computing pose-invariant feature descriptors of the image. The system and method also allows for the construction of three-dimensional (3D) object and class models from the pose-invariant interest points and feature descriptors of previously obtained scenes. Interest points and feature descriptors of a newly acquired scene may be compared to the object and/or class models to identify the presence of an object or member of the class in the new scene.

Abstract: A system for obtaining a probable 3D map of an occluded surface of an object is provided. The system receives an initial 3D map of a visible surface of the object and identifies one or more symmetries of the initial 3D map. The system computes the probable 3D map of the occluded surface by projecting points of the initial 3D map into occluded regions of space according to the identified symmetries. The system can also comprise an imager for obtaining the initial 3D map. The actual occluded surface cannot be known with absolute certainly because it is occluded; however, the computed 3D map will closely resemble the actual occluded surface in many instances because most objects possess one or more symmetries and the computed 3D map is based on such symmetries that are identified in the initial 3D map of the object.

Abstract: A massively parallel computer system is disclosed having a global router network in which pipeline registers are spatially distributed to increase the messaging speed of the global router network. The global router network includes an expansion tap for processor to I/O messaging so that I/O messaging bandwidth matches interprocessor messaging bandwidth. A route-opening message packet includes protocol bits which are treated homogeneously with steering bits. The route-opening packet further includes redundant address bits for imparting a multiple-crossbars personality to router chips within the global router network. A structure and method for spatially supporting the processors of the massively parallel system and the global router network are also disclosed.

Abstract: A massively parallel computer system is disclosed having a global router network in which pipeline registers are spatially distributed to increase the messaging speed of the global router network. The global router network includes an expansion tap for processor to I/O messaging so that I/O messaging bandwidth matches interprocessor messaging bandwidth. A route-opening message packet includes protocol bits which are treated homogeneously with steering bits. The route-opening packet further includes redundant address bits for imparting a multiple-crossbars personality to router chips within the global router network. A structure and method for spatially supporting the processors of the massively parallel system and the global router network are also disclosed.