Skip to content

tkyoung13/EviSeg

Repository files navigation

EviSeg Tutorial

Calculation of the conflict

The further insights and implementation details are provided in these python files.The following will illustrate how to compute the conflict value using a semantic segmentation model as an example.

Installation

# Get Semantic Segmentation source code
git clone https://github.com/tkyoung13/Calculation-of-the-conflict.git

Before running the code, we recommend install the following requirements:

  • An NVIDIA GPU and CUDA 9.0 or higher. Some operations only have gpu implementation.
  • PyTorch (>= 0.5.1)
  • numpy
  • sklearn
  • scikit-image
  • pillow
  • tqdm
  • tensorboardX
  • opencv-python
  • apex

Network architectures

Here we use DeepLabV3+ architecture with WideResNet38 backbones. There are other options including SEResNeXt(50, 101) and ResNet(50,101).

Pre-trained models

Pre-trained models have been provided. Please download the checkpoints to a designated folder pretrained_models.

Conflict calculation demo for a folder of images

If you want to try trained model on Camvid datasets, simply use

bash scripts/eval_camvid.sh pretrained_models/camvid_best.pth results/ --save-dir YOUR_SAVE_DIR

This snapshot is trained on CamVid dataset, with DeepLabV3+ architecture and WideResNet38 backbone. The predicted conflict images will be saved to YOUR_SAVE_DIR. Check it out.

Acknowledgments

Parts of the code were heavily derived from Improving Semantic Segmentation via Video Prediction and Label Relaxation.

About

Uncertainty Quantification Method for Semantic Segmentation Model

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published