DANCE: Dual Adaptive Noise-Cancellation and Enhancement for One-Dimensional Signals

🚀 Quick Start

Follow these steps to set up the environment and run the model.

1. Installation

Prerequisite: Ensure your Python version is 3.10.

We recommend creating a virtual environment (e.g., using Conda) to manage dependencies:

conda create -n dance python=3.10
conda activate dance

Then, install the necessary dependencies:

pip install -r requirements.txt

2. Data Preparation

Please download the following datasets from PhysioNet:

• MIT-BIH Arrhythmia Database: https://physionet.org/content/mitdb/
• MIT-BIH Noise Stress Test Database: https://physionet.org/content/nstdb/

📂 Directory Structure

After downloading and unzipping, place the dataset folders in the top-level directory (DANCE/). Your project structure should look like the tree below:

DANCE/
├── mit-bih-arrhythmia-database/         <-- Place Dataset 1 here
├── mit-bih-noise-stress-test-database/  <-- Place Dataset 2 here
├── datasets/
│   ├── split_manager.py
│   └── ...
├── script/
│   └── DANCER.sh
├── requirements.txt
└── README.md

Note: Ensure the folder names match the structure above so the scripts can locate the files correctly.

3. Preprocessing

Run the data manager script to preprocess and split the data:

python datasets/split_manager.py

4. Training & Usage

To run the model, execute the provided shell script:

bash script/DANCER.sh

Alternatively, you can modify script/DANCER.sh to customize parameters for your specific instances.

---

Dan Liu, Tianhai Xie @ IIP-2025