Development and Adaptation of Robotic Vision in the Real–World: the Challenge of Door Detection

@misc{antonazzi2024development,
      title={Development and Adaptation of Robotic Vision in the Real-World: the Challenge of Door Detection}, 
      author={Michele Antonazzi and Matteo Luperto and N. Alberto Borghese and Nicola Basilico},
      year={2024},
}

Here you can find the code and the datasets of the article entitled Development and Adaptation of Robotic Vision in the Real–World: the Challenge of Door Detection link

Installation: pip3 install -e .

Examples inferece code:

• YOLOv5 here
• Faster R-CNN here
• DETR here

Code:

• The code related to the position extraction algorithm is contained inside positions_extractor package.
• The source code of the object detectors (DETR, YOLOv5, and Faster R-CNN) used in the experimental campaign can be found in the here
• The scripts to perform the training and the evaluation of the models are in the train_models and calculate results packages respectively.

Usefull links:

• The DeepDoors2 dataset link
• The simulated dataset collected with iGibson link
• The photorealistic simulated dataset collected with Gibson link
• The real dataset acquired with the Giraff-X robotic platform in 4 real environments link
• The parameters of the trained models can be downloaded here