This repository contains code for the blog post Convolutional Neural Network, CNN based Image Colorization using OpenCV.

The small support files (pts_in_hull.npy and models/colorization_deploy_v2.prototxt) are now stored directly in this repo. Run getModels.sh only to download the large colorization_release_v2.caffemodel from the repository release assets.

Attribution for the vendored support files and the mirrored model asset is documented in ASSET_ATTRIBUTION.md.

sudo chmod a+x getModels.sh ./getModels.sh 

Python: Commandline usage to colorize a single image: python3 colorizeImage.py --input greyscaleImage.png a video file: python3 colorizeVideo.py --input greyscaleVideo.mp4

C++:

g++ pkg-config --cflags --libs opencv4 colorizeImage.cpp -o colorizeImage.out -lopencv_core -lopencv_dnn -lopencv_highgui -lopencv_imgcodecs -lopencv_imgproc -std=c++11 g++ pkg-config --cflags --libs opencv4 colorizeVideo.cpp -o colorizeVideo.out -lopencv_core -lopencv_dnn -lopencv_highgui -lopencv_imgcodecs -lopencv_imgproc -lopencv_videoio -std=c++11

a single image: ./colorizeImage.out greyscaleImage.png a video file: ./colorizeVideo.out greyscaleVideo.mp4

Want to become an expert in AI? AI Courses by OpenCV is a great place to start.