README.md

RoReg Benchmark

Preprocessing (backbone extraction)

The first thing to do is to use run_preprocessing_global.py to do the preprocessing using FCGF.

RoReg run

Use run_benchamrk_global.py to run RoReg on the preprocessed data

Evaluate

Use evaluate_all.py to create the results folder with the results files

Evaluation on the "original" (RoReg paper) ETH experiments

Downalod the ETH dataset from https://drive.google.com/file/d/1hyurp5EOzvWGFB0kOl5Qylx1xGelpxaQ/view?usp=sharing. Place the downloaded ETH data in the data folder like this:

data/
├── origin_data/
    └── ETH/

Run the preprocessing with python3 testset.py

Then run the registration and evaluation with python3 Test.py --RD --RM --ET yohoo --keynum 1000 --testset ETH --tau_2 0.2 --tau_3 0.5 --ransac_ird 0.5