Year: 2017
Type: article
Abstract: Feature matching is one of the fundamental issues in computer vision. The established methods, however, do\nnot provide reliable results, especially for extreme viewpoint changes. Different approaches have been proposed to\nlower this hurdle, e. g., by randomly sampling different viewpoints to obtain better results. However, these methods\nare computationally intensive.\nIn this paper, we propose an algorithm to enhance image matching under the assumption that an image, taken in\nman-made environments, typically contains planar, rectangular objects. We use line segments to identify image\npatches and compute a homography which unwraps the perspective distortion for each patch. The unwrapped\nimage patches are used to detect, describe and match SIFT features.\nWe evaluate our results on a series of slanted views of a magazine and augmented reality markers. Our results\ndemonstrate, that the proposed algorithm performs well for strong perspective distortions.
Authors: Marco Filax, Tim Gonschorek, Frank Ortmeier
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3031722356