RefineContourNet

RefineContourNet (RCN) is an Object Contour and Edge Detector. This is the test code and the best results of the paper Object Contour and Edge Detection with RefineContourNet; CAIP 2018

DOI: 10.1007/978-3-030-29888-3_20

To test the RCN - please download refinenet-contour-master

compile and setup the included MatConvNet-files (same procedure like for RefineNet [https://github.com/guosheng/refinenet])
run the file demo_test_simple_voc_contour_tiff.m with MATLAB
choose in this m-file the trained model for either Object Contour Detection or Edge Detection

Best RCN-results:

RCNVOC_NMS_results_on_BSDS500_val
RCNVOC_results_on_BSDS500_val
RCN_NMS_results_on_PASCAL_val
RCN_results_on_PASCAL_val

About

Results of the paper "Object Contour and Edge Detection with RefineContourNet"

Other

Languages

Language:MATLAB 44.6%Language:Cuda 27.4%Language:C++ 11.7%Language:TeX 7.4%Language:Python 4.8%Language:Shell 1.6%Language:C 1.0%Language:Makefile 0.8%Language:HTML 0.3%Language:CSS 0.3%Language:M 0.1%Language:JavaScript 0.0%