/active-learning

This is the implementation code for the paper, "A semi-supervised self-training method to develop assistive intelligence for segmenting multiclass bridge elements from inspection videos"

Primary LanguageJupyter Notebook

Bridge element segmentation from inspection videos

This is the implementation code for the paper, "A semi-supervised self-training method to develop assistive intelligence for segmenting multiclass bridge elements from inspection videos".

ezgif com-gif-maker (4) ezgif com-gif-maker (1)

Getting Started

  • Install the required dependencies: (for reference see how_to_install.pdf )
  • Dataset: Download the initial dataset from here.
  • weights: Download pre-trained resnet_50 coco weights and trained weights for bridge element segmentation from here.
  • custom.py : this code is used for loading data and training the model
  • Training.ipynb: loading the weight and calling the training function
  • inference.ipynb: this code is used for inferencing.
  • mrcnn/visualize.py : this code is used for visualizing the segmented bridge elements with mask.

Citation

If you use this repository, please cite the following paper:

@article{karim2021semi,
  title={A semi-supervised self-training method to develop assistive intelligence for segmenting multiclass bridge elements from inspection videos},
  author={Karim, Muhammad Monjurul and Qin, Ruwen and Chen, Genda and Yin, Zhaozheng},
  journal={Structural Health Monitoring},
  pages={14759217211010422},
  year={2021},
  publisher={SAGE Publications Sage UK: London, England}
}

Note that part of the codes are referred from Mask RCNN project.