Paper Title: Deep Learning Approach for Accurate Segmentation of Sand Boils in Levee Systems

This repository contains codebase and links for datasets of our paper based on controlled transfer learning.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Prerequisites

You would need to install the following software before replicating this framework in your local or server machine.

Python version 3.7+
Aanaconda version 3+
TensorFlow version 2.12.0
Keras version 2.12.0

Download and install code

• Retrieve the code

git clone https://github.com/manisa/SandBoilNet.git
cd SandBoilNet

• Create and activate the virtual environment with python dependendencies.

conda create -n gpu-tf tensorflow-gpu
conda activate gpu-tf
pip install tensorflow==2.12.*

Download datasets

• Original Training Data
• Original Test Data
• Augmented Train and Validation Data
• Unzip and copy dataset from the respecitve experiment into the folder datasets inside the root folder SandBoilNet.

Download trained models

• All IEEE Access Models
• Unzip and copy models from respective experiment to models inside the root folder SandBoilNet.

Folder Structure

SandBoilNet/
    archs/
    lib/
    datasets/
        sandboil_augmented_5_8_23_6853/
        test/
    models/
        IEEE_models/
            Baseline_Conv_bce_dice_loss/
            Baseline_LeakyRI_bce_dice_loss/
            baseline_normal_bce_dice_loss/
            Baseline_ProposedAtt_bce_dice_loss/
            SandBoilNet/
            unet_bce_dice_loss/

Training

• To replicate the training procedure, follow following command line.

cd src
python train.py

Authors

MANISHA PANTA, KENDALL N. NILES, JOE TOM, MD TAMJIDUL HOQUE, MAHDI ABDELGUERFI AND MAIK FALANAGIN

License

This project is licensed under the MIT License - see the LICENSE.md file for details