U.S.-Italy Cooperative Research: High-Performance Parallel Computing for Computer Vision

9300146戴尔该奖项将支持威斯康星大学的查尔斯·R·戴尔和意大利帕维亚大学的马尔科·费雷蒂之间的合作研究。这项研究将尝试利用并行编程模型来克服形状和运动分析方面的现有问题。它将包括为高性能并行计算机设计和实现并行计算机视觉算法的研究。将为二维和三维形状分析以及运动分析的中级视觉任务开发数据并行方法。每个研究人员都为这一合作努力带来了互补的优势。戴尔在三维物体表示、物体识别、运动分析以及使用连续、定向观察者运动的主动视觉三维形状恢复和探索方法领域开发了许多新技术。Ferretti广泛研究了并行算法在各种定制和通用并行计算机上的实现。研究人员将在确定的共同感兴趣的问题上进行合作，这将增强两组开发新技术和实施并行程序测试这些方法的能力。***

9300146 Dyer This award will support collaborative research between Charles R. Dyer, University of Wisconsin, and Marco Ferretti, University of Pavia, Italy. The research will attempt to exploit parallel programming models to overcome existing problems in shape and motion analysis. It will include research on designing and implementing parallel computer vision algorithms for high-performance parallel computers. Data-parallel methods will be developed for the intermediate-level vision tasks of two- dimensional and three-dimensional shape analysis, and motion analysis. Each researcher brings complementary strengths to this cooperative effort. Dyer has developed a number of new techniques in the areas of three- dimensional object representation, object recognition, motion analysis, and methods for active visual three-dimensional shape recovery and exploration using continuous, directed observer motion. Ferretti has studied extensively the implementation of parallel algorithms on a wide range of custom and general-purpose parallel computers. The researchers will collaborate on defined problems of mutual interest that will enhance both groups' abilities to develop new techniques and implement parallel programs to test these methods. ***