Beck-Sisyphus/intelligent_scissor

computer vision introduction practice on image segmentation

intelligent_scissor

computer vision introduction practice on image segmentation

Feb. 28rd, 2018

PAN, Jie && PANG, Sui

Build

mkdir build
cd build
cmake ..
make
./intelligent_scissor

Supported Environment: macOS High Sierra Ubuntu 16.04 LTS

Toolchain: editor: CLion language: C++ compiler: CMake/make dependency: OpenCV 3.2 GUI: Qt

Core Algorithm

1. Convert the imported picture to cost diagram from its derivative.
2. Store the pixels as nodes, with the following field:

class Pixel_Node : public FibHeapNode
{
public:
    int state;
    int row, col;

    int link_cost[9];
    long total_cost;

    Pixel_Node* prevNode; // connecting to multiple other nodes called graph
};

3. A priority queue implemented by fibonacci heap to store and process the nodes
4. A Dijkstra's algorithm for calculating the minimum path

User Interface

We also needs to handle the user interface well for finishing this project. There are three criterias:

1. Handling different type of images, jpeg, png, gif, etc.

2. Handling mouse and keyboard command, as listed here:

   • Ctrl+"+", zoom in;

   • Ctrl+"-", zoom out;

   • Ctrl+Left click first seed;

   • Left click, following seeds;

   • Enter, finish the current contour;

   • Ctrl+Enter, finish the current contour as closed;

   • Backspace, when scissoring, delete the last seed; otherwise, delete selected contour.

3. Store and go in the debug mode.

Artifact

We bring our most fashionable lady on to our most fashionable chassis.