Year: 2008
Type: conference-paper
Abstract: In computer vision applications, it is important to extract and describe robust and characteristic features from objects in images.Many recognition and tracking algorithms rely on this information.One of the main problems with current state-of-the-art algorithms, such as SIFT [1] , GLOH [2] , SURF [3], is that the retrieval of features at the border of an object is sensitive to changes in the background.This work presents a possible solution using background/foreground segmentation
Author: Dimitri Van Cauwelaert
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2300110852