This page presents the downloadable results for the paper

“Two-view line matching algorithm based on context and appearance”

Juan López , Roi Santos, , Xosé R. Fdez-Vidal , Xosé M. Pardo

http://www.sciencedirect.com/science/article/pii/S0031320314004889

GitLab repository: https://gitlab.com/rsan/line/tree/36df922f523219afd15f8a541e59980992b380f9

The results are classified in the following directories:

Directory name: executablesWin64

detect.exe : Performs detection and plotting of detected lines.

Input: jpg/png image file.

Output: xxx_RESULT_detect.txt text file with the list of detected lines; each line is described by both endpoints, length and orientation. Image with plotted detected lines.

Example: detect 04-1.jpg

match.exe : Performs detection, matching and plotting of matched lines

Input: Two jpg/png image file

Output: Text file with the list of matched lines; each line is described by both endpoints on the original image, and its length and orientation. Images with plotted matched lines, labeded with matching reference tag.

Example: match 04-1.jpg 04-2.jpg

- Images database: Classified image database. Please check the references:

The directories matching_industriales, detection_averageGray, detection_real_lum, detection_real_noise, detection_synthetic_lum, detection_synthetic_noise contain the images selected for the paper.

Reference: Two-view line matching algorithm based on context and appearance in low-textured images. Pattern Recognition Volume 48, Issue 7, July 2015, Pages 2164–2184 http://www.sciencedirect.com/science/article/pii/S0031320314004889

The directory matching_publicas contains images from public databases. Their references are:

http://www.vision.ee.ethz.ch/ZuBuD

http://www.robots.ox.ac.uk/

http://lear.inrialpes.fr/people/mikolajczyk/Database/index.html

  • roisantos/hycnn.txt
  • Última modificación: 2016/06/01 10:24
  • por roi.santos