On the Applicability of Synthetic Data for Re-Identification in Warehousing Logistics

This repository complements our paper proposal "On the Applicability of Synthetic Data for Re-Identification in Warehousing Logistics".

This work is part of the project "Silicon Economy Logistics Ecosystem" which is funded by the German Federal Ministry of Transport and Digital Infrastructure.

Example results

C2RL

row 1: Centered pallet block -> Rotated left pallet block -> Reconstructed centered pallet block

Usage

• Environment

  • Python 3.6

  • TensorFlow 2.2, TensorFlow Addons 0.10.0

  • OpenCV, scikit-image, tqdm, oyaml

  • we recommend Anaconda or Miniconda, then you can create the TensorFlow 2.2 environment with commands below

    conda create -n tensorflow-2.2 python=3.6
    
    source activate tensorflow-2.2
    
    conda install scikit-image tqdm tensorflow-gpu=2.2
    
    conda install -c conda-forge oyaml
    
    pip install tensorflow-addons==0.10.0

  • NOTICE: if you create a new conda environment, remember to activate it before any other command

    source activate tensorflow-2.2

• Dataset

  • download the pallet-block-502 dataset and extract the images you need

    https://zenodo.org/record/6353714

  • or take the filtered part of the dataset that we used from here

    https://zenodo.org/record/6580127

• Example of training

  CUDA_VISIBLE_DEVICES=0 python train.py --dataset pallet-block-502

  • tensorboard for loss visualization

    tensorboard --logdir ./output/pallet-block-502/summaries --port 6006

• Example of testing

  • To generate images using the trained cycleGAN

  CUDA_VISIBLE_DEVICES=0 python test.py --experiment_dir ./output/pallet-block-502

• The checkpoints for the CycleGAN trained on pallet-block-502, the classifier model as well as the output images of the GAN can be downloaded here

      https://zenodo.org/record/6580127

  • The downloaded weights should be placed in ./output/pallet-block-502/checkpoints/
  • The downloaded classifier model (and json file) should be placed in ./model_classifier_C_RL/

• Download the pallet block dataset and load the cycle GAN checkpoints along with the weights of the classifier

      sh download_pallet_dataset_and_load_weights.sh