Color Control Examples
Images used:
https://github.com/leongatys/NeuralImageSynthesis/blob/master/Images/ControlPaper/fig3_style1.jpg
https://github.com/leongatys/NeuralImageSynthesis/blob/master/Images/ControlPaper/fig3_content.jpg
The style image is adjusted to match the content image:
After Neural-Style:
The final output:
The process only takes 4 commands (including neural_style.lua) to complete:
Step 0 (Optional Depending On Your Content/Style Image):
python linear-color-transfer.py --target_image fig3_style1.jpg --source_image fig3_content.jpg --output_image style_colored_pca.png
Step 1:
python lum-transfer.py --content_image fig3_content.jpg --style_image style_colored_pca.png --cp_mode lum --output_style_image output_lum_style_pca.png --output_content_image output_lum_content_pca.png --org_content fig3_content.jpg
Step 2:
th neural_style.lua -original_colors 0 -image_size 1000 -content_weight 1 -style_weight 1e3 -output_image out_lum6_test.png -content_image output_lum_content_pca.png -style_image output_lum_style_pca.png -num_iterations 1500 -save_iter 50 -print_iter 50 -seed 876 -init image -backend cudnn -cudnn_autotune
Step 3:
python lum-transfer.py --output_lum2 out_lum6_test_400.png --cp_mode lum2 --output_image out_combined.png --org_content fig3_content.jpg
Step 4 (Optional):
python linear-color-transfer.py --target_image out_combined.png --source_image fig3_content.jpg --output_image out_combined_pca.png