X-PDNet: Accurate Joint Plane Instance Segmentation and Monocular Depth Estimation with Cross-Task Distillation and Boundary Correction (BMVC 2023)

This is an implementation for X-PDNet: a multi-task learning framework for joint plane instance segmentation and depth estimation

The official paper can be found at paper. Thank the PlaneRecNet for a great baseline implementation

How to run the inference?

1. Use conda to create an env:

conda env create -f environment.yml

2. Create a folder "weights", download resnet and X-PDNet checkpoint via this link and put on "weights" folder.
3. Inference a single image (*.png or *.jpg for mat):

python3 simple_inference.py --config=XPDNet_101_config --trained_model=weights/XPDNet_101_9_125000.pth  --image=example
_images/scene0134_01_frame_color_756.jpg

4. Inference a folder:

python3 simple_inference.py --config=XPDNet_101_config --trained_model=weights/XPDNet_101_9_125000.pth --images=input_folder:output_folder