Year: 2009
Type: conference-paper
Abstract: A large number of processes in computer vision are based on the image motion measurement, which is the projection of the \nreal displacement on the focal plane. Such a motion is currently approximated by the visual displacement field, called optical \nflow. Nowadays, a lot of different methods are commonly used to estimate it, but a good trade-off between execution time and \naccuracy is hard to achieve with standard integrations. This paper tackles the problem by proposing a parallel implementation \nof the well-known pyramidal algorithm of Lucas & Kanade, in a Graphics Processing Unit (GPU). It is programmed using the \nCompute Unified Device Architecture from NVIDIA corporation, to compute a dense and accurate velocity field at about 15 Hz \nwith a 640 480 image definition.
Authors: Marzat, Julien, Dumortier, Yann, Ducrot, André
Language: English
Topic: Advanced Vision and Imaging
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2531394286