An integrated system for 3D motion estimation and object recognition with outdoor stereo image sequences as inputs is presented. The goals are to obtain from input stereo images the 3D motion description and the identification of a vehicle moving against a stationary background. In order to accomplish the desired goals, the system consists of four stages; motion estimation, distinctive feature extraction, model database, and object recognition. The results of 3D motion estimation are used to narrow down considerably the search space in the stage of object recognition. The system is applied to eight sets of stereo image pairs. Experimental results demonstrate that the system can successfully provide the 3D motion description and the identification of the vehicle in all the stereo image pairs.>

Several techniques to perform static and dynamic load balancing for vision systems are presented. These techniques capture the computational requirements of a task by examining the data when it is produced. They can be applied to many vision systems because many algorithms in different systems are either the same or have similar computational characteristics. These techniques are evaluated by applying them on a parallel implementation of the algorithms in a motion estimation system on a hypercube multiprocessor system. It is shown that the performance gains when these data decomposition and load balancing techniques are used are significant and that the overhead of using these techniques is minimal.>

A method for detecting the wheels of a vehicle in stereo image pairs is presented. The method consists of two steps: geometrical transformation and circle extraction. The geometrical transformation uses the disparity values obtained from a stereo image pair to calculate the parameters of the plane containing wheels of the vehicle. These parameters are used to transform any elliptical wheels contained in the plane to circular ones which can be extracted by the circle extraction algorithm. The circle extraction algorithm consists of template matching and the Hough transform. In order to save computation and improve the results in the Hough transform, two constraints, the neighbor-region edge connectivity and the gradient direction of each edge point, are used to eliminate noncircular edge points. Experimental results show that these two constraints do eliminate noncircular edge points and preserve any circle embedded in edges. The final results show that the proposed method can detect and locate the wheels of a vehicle successfully.>