Image colorization using the Colorization using Optimization method DOI:10.1145/1015706.1015780
this project is mainly a cli tool rn, but i am planning to publish the underlying colorization library too. i have implemented this algorithm in python already inside of a GIMP plugin that i made. and although the speed of the colorization has dramatically improved after rewriting a critical section of the algorithm in C, it can still reach more than 120 seconds for large images. im hoping a complete rewrite in c++ could be way more performant. this is NOT a GIMP plugin ...yet
- handling and evaluating cli arguments
- colorizing a grayscale image based on colored marks/tags/hints drawn onto it
- option to select a different affinity (weights calculation) function (a set predefined functions, not custom (unless you modify the source code))
- verbose option
- multithreading
- option to draw the marks when running the command using a gui (qt?) instead of specifying a path to the tagged image
- option to immediately show the colored image after the colorization is done in an image viewer (feh, loupe, ...)