VCells is the implementation of VCells algorithms described in VCells: Simple and Efficient Superpixels Using Edge-Weighted Centroidal Voronoi Tessellations, TPAMI 34(6), 1241-1247" by Jie Wang and Xiaoqiang Wang.
This project has 2 components:
VCells
├── VCellsCpp
├── VCellsServer
It adopts the code of main algorithm in original implementation: VCells.cpp. We've replaced the image reading and writing part with OpenCV's component, so now it can support most image format. Besides, it now supports command line to specify which image to operate.
In order to provide a friendly user interface, we build a simple web application, where Golang is used to build a simple server and jQuery and Bootstrap are used to design the interface.
We use CMake in our project, so you have to generate makefile with CMake and then build target with makefile.
We use OpenCV in our project. So, you should build and install OpenCV on your machine. You can also download a release of VCellsCpp on our Release Page.
cd VCellsCpp
mkdir cmake-build-debug && cd cmake-build-debug
cmake ..
You will see a makefile generated in directory cmake-build-debug.
make
This will generate an executable file.
cd VCellsServer
go build main.go file_upload.go
And you will get an executable file: main
You can use the VCellsCpp part independently as a command line program. Or, you can run a simple server on your machine and interact with it from web browser.
VCellsCpp serves as a command line program. Its usage is as follows:
./VCellsCpp [srcImgPath]
It will generate a initial segmented image and a superpixel image in the same directory of source image.
Before you run the server, you should copy the executable VCellsCpp to the directory VCellsServer/upload_files/.
You can specify the port that the server listens to and the log file path. The default port will be set to 8080,
and the default log file path will be set to ./vcells.log.
./main -port [port] -log [log]