ASJ is a novel compact geometric feature for characterising images.
If you use this code for research please cite:
@article{XXBZS18,
author = {Xue, Nan and Xia, Gui-Song and Bai, Xiang and Zhang, Liangpei and Shen, Weiming},
journal = {IEEE Transactions on Image Processing},
title = {{Anisotropic-Scale Junction Detection and Matching for Indoor Images.}},
volume = {27},
number = {1},
pages = {78-91},
year = {2018}
}
Building and using requires the following libraries and programs
OpenCV
The code has been tested on Ubuntu 14.04 and Windows 10.
git clone [email protected]:cherubicXN/anisotropic-scale-junction-detector.git
cd anisotropic-scale-junction-detector
mkdir build
cd build
cmake ..
make
Linux:
./ASJDetector ../example_image/im1.jpg
Windows:
ASJDetector.exe ../example_image/im1.jpg
The first line of *.asj will contain a number
location_x location_y
junctionClass scale rd logNFA
branch_0 strength_0 scale_0
branch_1 strength_1 scale_1