GPUVision is a framework for Image Processing, Computer and Robot Vision exploiting the GPU for its parallel processing capabilities. There are a number of common utilities implemented which can be copied/modified and combined to create new specialized processing modules. This was started at UPenn as a final project by Ikkjin Ahn and Michael Lehr for CIS 700-010, GPU Programming and Architecture, taught by Suresh Venkatasubramanian from AT&T Labs. GPUVision is being developed as OpenSource under the GPL License.   Ikkjin Ahn, Michael Lehr http://gpuvision.sourceforg.net/ gpuvision@gmail.com July 1, 2005