The OpenVIDIA project implements computer vision algorithms on computer graphics hardware, using OpenGL and Cg. The project provides useful example programs which run real time computer vision algorithms on single or parallel graphics processing units(GPU).
This project is released as part of the work conducted at the Eyetap Personal Imaging Lab (ePi Lab) at the Electrical and Computer Engineering Group at the University of Toronto.
OpenVIDIA utilizes the computational power of the GPU to provide real--time computer vision much faster than the CPU is capable of, and leaves the CPU free to conduct other tasks beyond vision.
This includes a program called "FLfeature" for recording and detecting scenes, and overlaying graphics from PNG files. Simply name your scene "myscene", and then create a PNG overlay, name it "myscene-overlay.png", and it will be loaded when the scene is deteected. 0.8.3 improves memory handling, is windows portable, and uses a single header file.
Chris Aimone (Vision Algorithms, 3d registrations)
Billal Belmellat (Cal3D rendering interface)
Daniel Chen (Handtracking algorithms)
Martin Dvh (Video4Linux support)
James Fung (Developer)
Ben Hur (low vision applications for the visually challenged)
Rosco Hill (3d windowing/UI interactions)
Mohit Kansal (Cal3D rendering interface),
Rick Lentz (windows port) Daniel Ng (low vision applications for the visually challenged)
Suryadi (3D studio model loader)
Felix Tang (Developer)
OpenVIDIA is currently being developed by volunteers and students, and new help is always welcome. This can range from contributing source code, bug fixes, or documentation. If you have a project related to OpenVIDIA, let us know, and we'd be happy to put a page link to it as well. Contact the project admin email openvidia [at] eyetap.org for more information.
OpenVIDIA Windows Support
2006-03-23 13:12 - OpenVIDIA : Parallel GPU Computer Vision
OpenVIDIA now compiles upon windows. Included is a package containing precompiled windows executables, from the 0.8.3 package source.
OpenVIDIA wins ACM Multimedia Open Source Software Award
2005-11-10 11:51 - OpenVIDIA : Parallel GPU Computer Vision
OpenVIDIA was chosen as this year's winner of the ACM Multimedia Open Source Software Competition. We'd like to thank ACM MM for this distinction.
OpenVIDIA version 0.8.0
2005-06-27 02:59 - OpenVIDIA : Parallel GPU Computer Vision
libopenvidia-0.8 is released. It is a stripped down but better organized/implemented version of openvidia. 0.8.0 includes simpler dc1394 capture, an easy to use feature tracker (much optimized for speed, and some bug fixes), and a c++ STL based feature handling API. Also, a simple GenericFilter class is implemented. There are 4 example programs showing how to use each of these, and class documentation is being included in the /docs directory and through the website. Eventually all of the functionality from 0.07x will be included into 0.8. 0.8.0 is sort of a pre-release if for anyone wanting some basic functions with an easy to use API. If there are features you'd like to see in 0.8 sooner than later just send an email so we know which features are a priority for users. Additionally we'd like to thank our sponsors, in particular, NVIDIA for their hardware contributions which greatly help this work.
OpenVIDIA version 0.07 Released
2004-12-14 20:24 - OpenVIDIA : Parallel GPU Computer Vision
OpenVIDIA v0.07 is released. OpenVIDIA provides a GPU accelerated computer vision framework built with OpenGL and Cg. This release focuses on an Open Source (GPL) image motion tracking framework for mediated / augmented / mixed reality environments. v0.07 includes a FLTK-gui based video tracker called flcornersTrack, allowing for easy attachment of virtual tags to live video. Notably, the 0.07 version deals with greater ranges of camera motion than before, particularly rotation about the optical axis. It is based on a feature point tracking method, and provides real-time feature correspondences which can be used for many higher level vision tasks. Additionally, video4linux2 support is included in an example program for those that need to use it. 3d studio models are loadable (.3ds files) into 3d coordinate systems in the 3dcoordCamFrustum program.
Openvidia 0.04 Released
2004-11-03 19:12 - OpenVIDIA : Parallel GPU Computer Vision
Openvidia-0.04 version includes the programs used to make the screenshots on the site. Included is a corner tracker (cornersTrack), which provides features and point correspondences, useful for higher level tracking tasks.
OpenVIDIA Initial Release
2004-05-12 22:08 - OpenVIDIA : Parallel GPU Computer Vision
We're happy to announce the initial code release of the OpenVIDIA project. OpenVIDIA is a project which seeks to create hardware accelerated computer mediated reality by using the Graphics Processing Unit (GPU) for computer vision.
OpenVIDIA is Free Software in the terms of the GPL. This means you have four basic freedoms:
The GPL is the legal mechanism that gives you these freedoms. It also protects them from being taken away: any derivative work based on the program must be under the GPL.
For those interested in commercial use under different terms, contact the project admins at openvidia [at] eyetap.org for information on dual licensing arrangements.
Transferring site with MySQL on PostgreSQL Part 1