Skip to content

Latest commit

 

History

History
40 lines (25 loc) · 2.83 KB

ONLINE_SEGMENTATION_EXP.md

File metadata and controls

40 lines (25 loc) · 2.83 KB

You can run the main experiments reported in the work Fast Object Segmentation Learning with Kernel-based Methods for Robotics by following the reported instructions.

Code installation

Please follow the instructions for installations reported at this link.

Datasets Download

To reproduce the experiments of the paper, you need to download the YCB-Video dataset. In order to have data in a standard format which is employed also by other robotic dataset, we download it from the BOP challenge website.

cd $HOME_DIR/Data/datasets/YCB-Video/test
wget http://ptak.felk.cvut.cz/6DB/public/bop_datasets/ycbv_test_all.zip
unzip ycbv_test_all.zip && mv test/* . && rmdir test && rm ycbv_test_all.zip

cd $HOME_DIR/Data/datasets/YCB-Video/train_real
wget http://ptak.felk.cvut.cz/6DB/public/bop_datasets/ycbv_train_real.zip
unzip ycbv_train_real.zip && mv train_real/* . && rmdir train_real && rm ycbv_train_real.zip

Note: at the end of the installation, remember to unset HOME_DIR

Usage

By modifying the configuration files in the experiments/configs folder and by substituting the files with the proper ones for your data, this code allows you to run customized experiments. However, in this repository we provide you with the configuration files and scripts that are required to reproduce the main experiment in the presented paper. To do this, you have to run the script experiments/run_experiment_segmentation.py and to properly set command line arguments. Some examples will be provided below. For additional information please refer to the helper of the script, running the command python run_experiment_segmentation.py -h in the experiments directory.

In the experiments/configs you can find two categories of configuration files related to this paper:

  • config_feature_extraction_segmentation_ycbv.yaml: sets parameters for feature extraction of Ours experiment, reported in the second row of Table 1 in the paper.
  • config_online_detection_segmentation_ycbv.yaml: sets parameters to train online detection and online segmentation of Ours experiment, reported in the second row of Table 1 in the paper.

Important: if you have more than one GPU available, before running an experiment, you have to set the number of the GPU that you want to use (only one) with the command export CUDA_VISIBLE_DEVICES=number_of_the_gpu

YCB-Video with COCO FEATURE-TASK experiment

To reproduce the results of Ours experiment in the second row of Table 1, in the experiments folder you have to run the command

python run_experiment_segmentation.py.

If you do not set from command line an output directory, experiment's results will be saved in the experiments/online_segmentation_experiment_ycbv/result.txt file.