Skip to content
A c++ implementation of pdollar's Structured Edge Detection Toolbox. Written by opencv c++.
C++ CMake
Branch: master
Clone or download
AlexMa011 Merge pull request #1 from zhuqiang00099/master
fix resize bug in edgeBoxes.cpp
Latest commit 288c4a6 Jun 7, 2018
Type Name Latest commit message Commit time
Failed to load latest commit information.
img add image folder Feb 6, 2018
include modify readme Feb 6, 2018
model delete unnecessray files Feb 6, 2018
output Merge branch 'master' of… Feb 6, 2018
src fix resize bug Jun 7, 2018
tools add edges result Feb 6, 2018
.gitignore modify readme Feb 6, 2018
CMakeLists.txt change some function names Feb 6, 2018 Update Feb 26, 2018
main.cpp fix some bugs Feb 6, 2018
parameters.cpp fix some bugs Feb 6, 2018


This is a c++ implementation of pdollar's Structured Edge Detection Toolbox.


Opencv2 is required. The installation documentation can be find here. (Note: opencv 3 is also plausible if the writeout function in gadgets.cpp is modified to suit opencv3.) And you need to install the matiolib(needed to load model in mat format). Download the matio package and install the matio library.

If you are using Mac OS with homebrew, just use brew install to install opencv2 and libmatio.

Getting Started

Switch to the main folder, and type the following command to run compile and run:

$ cmake .
$ make
$ ./edgebox

Remember to delete CMakeCache.txt if needed.

If you want to see the pic immediately, set the showpic para to be 1. The result image is stored in output folder.

Tools folder contains function for image processing. The parameters.cpp file contains function about loading model, type conversion and initialize parameters. Parameters can be changed in initial_para() function in parameters.cpp.


Origin image:


Edge image:


Detection image:


You can’t perform that action at this time.