LKRM

This repo is the implementation of the paper ("Latent Knowledge Reasoning Incorporated for Multi-fitting Decoupling Detection on Electric Transmission Line"). The code is based on PyTorch and large part code is reference from faster-rcnn.

Requirements

• Python3.8
• Python packages
  • PyTorch >= 1.0
  • Torchvision >= 0.9.0
  • opencv-python
  • scipy
  • matplotlib
  • numpy

Demo

After successfully completing requirements, you can be ready to run the demo.

• Download the cascade_fpn_1_12_2325.pth which finally use in the paper(LKRM) from Weights (extract code:idfv)

• Download the pretrained weights(pascal_voc_cascade.pth and resnet101_caffe.pth) from Weights (extract code:idfv)

• Put cascade_fpn_1_12_2325.pth into the:

{repo_root}/models/res101/pascal_voc/0.0018_9_0.1_023010/

• Put pascal_voc_cascade.pth into the:

{repo_root}/models/

• Put resnet101_caffe.pth into the:

{repo_root}/data/pretrained_model/

• Using this code to see the fitting detection results in demo images:

python cascade_test_net.py --cuda

Citation

@article{jjfaster2rcnn,
    Author = {Jianwei Yang and Jiasen Lu and Dhruv Batra and Devi Parikh},
    Title = {A Faster Pytorch Implementation of Faster R-CNN},
    Journal = {https://github.com/jwyang/faster-rcnn.pytorch},
    Year = {2017}
}

@inproceedings{renNIPS15fasterrcnn,
    Author = {Shaoqing Ren and Kaiming He and Ross Girshick and Jian Sun},
    Title = {Faster {R-CNN}: Towards Real-Time Object Detection
             with Region Proposal Networks},
    Booktitle = {Advances in Neural Information Processing Systems ({NIPS})},
    Year = {2015}
}