Including Chamfer-L2 Distance, F-score and Normal Consistency Score.
Run the script:
cd vanilla_metric
python eval_two_folder.py --eval_type syn_obj --in_dir DIR_to_THE_INPUT --gt_dir DIR_to_GROUND_TRUTH --samplepoints 200000 --out_csv test.csv
The results could be recorded in test.csv
Use the pretrained model to evaluate
cd vanilla_metric
python eval_two_folder.py --eval_type syn_obj --in_dir DIR_to_THE_INPUT --gt_dir DIR_to_GROUND_TRUTH --model_dir DIR_to_MODEL --out_csv test.csv
Train the network
cd vanilla_metric
python Train.py --name MODEL_NAME_to_SAVE
The results could be recorded in test.csv