Content Aware Platte-based Image Recoloring

The source code of the paper “Palette-based Content-Aware Image Recoloring”

Requirements

1. OS: Windows 10
2. Qt: 5.12.9
3. OpenCV: 4.1.2
4. IDE: Visual Studio 2019
5. OpenGL

Usage

1. Click RgbPalette_Recolor_GUI-1.sln to compile the source code
2. Click the button Open Image & Semantic map to load the input image and semantic features (see the directory "data", semantic feature data named as "xxx_feat.data"), then the original and recolored images will be shown on the right
3. Fill the palette size and Click the button Extract Palette to extract the color palette of the input image
4. Click the button Calc. Weight to calculate the mixing weights
5. Modify the bellow palette colors to recolor the input images

Semantic features extraction

1. We follow Yagiz Aksoy's code to extract the semantic features, please refer to https://github.com/iyah4888/SIGGRAPH18SSS and replace the main_hyper.py with SemanticExtractor.py
2. Run SemanticExtractor.py to extract the semantic feature of an input image (could run with CPU)

Contact info

• Email Addr.: dzj@qhu.edu.cn
• Paper: https://zhengjun-du.github.io/papers/202312/content-aware%20recolor.pdf