FastGraphCut

Final Project for How to Write Fast Code II (18-646) at Carnegie Mellon University.

This project implements a parallelized version of the GrabCut image segmentation algorithm, aiming to improve runtime performance over a sequential baseline.

Requirements

• OpenCV (version 4.x recommended)
• CMake (version 3.10 or higher)
• A C++17 compatible compiler (e.g., g++, clang++)

Build

• mkdir build
• cd build
• cmake ..

Running sequential version

• cd build
• make
• ./SlowGrabCut <image_path> x1 y1 x2 y2

Where image_path is the path to your image, and x1, y1, x2, y2 are the coordinates of the bounding box.

e.g. ./SlowGrabCut ../dataset/large/flower.jpg 531 300 3383 2101

Running parallelized version

• cd build
• make
• ./FastGrabCut <image_path> x1 y1 x2 y2

Where image_path is the path to your image, and x1, y1, x2, y2 are the coordinates of the bounding box.

e.g. ./FastGrabCut ../dataset/large/flower.jpg 531 300 3383 2101