Local-Laplacian-Filters

Reproduction of the paper "Local Laplacian filters: edge-aware image processing with a Laplacian pyramid" for the course Advanced Digital Image Processing at TU Delft

How to run the algorithm?

Requirements:

• NumPy
• OpenCV
• PIL
• SharredArray

Options:

• img_path: path to the input image
• out_path: path to save the results
• color_img: Indicates whether we process color or grayscale image.
• intensity_img: Whether to use intensity image for processing color image
• mapping_func: Type of remapping function 'color' or 'grayscale'
• levels: Number of levels for Gaussian/Laplacian Pyramid
• sigma: Algorithm Hyperparameter
• alpha: Algorithm Hyperparameter
• beta: Algorithm Hyperparameter
• num_processes: Number of processes to run the algorithm

Example Usage:

python main.py --img_path data/desk_256.hdr --out_path results/result.png --color_img True --intensity_img True --mapping_func grayscale --levels 5 --sigma 0.4 --alpha 1.0 --beta 0.5 --num_processes 16

Exemplary results (Detail manipulation):