Reference: Anat Levin's colorization code
See: www.cs.huji.ac.il/~yweiss/Colorization/
author : Hyungtae Lim (shapelim@kaist.ac.kr)
This code is for generating full depth using sparse depth and RGB image,based on taskonomy.
Taskonomy dataset maxdepth is 128m,data processing is depth_png/(2**16 -1) * 128
The sparse depth is 16bit
If you want to fill in a sparse depth map, and visualize(Note that there is no depth map procedure stored)
$ python3 main.py
If you want to fill multiple sparse depth maps, and save them to 16 bits
$ python3 main_images.py
- If you run this code by your own sparse depth and RGB image, then revise sparse_depth_path and rgb_path on main.py
- It takes some time for optimization