Digital Twin of Lung from Wearable Biosignals to Real-Time Monitoring

Partha Acharya¹, Nikhileswara Rao Sulake², Soutrik Chakraborty², Subhamoy Mandal², Suman Chakraborty³

¹School of Medical Science and Technology, Indian Institute of Kharagpur, India
²Department of Computer Science, RGUKT Nuzvid
³Department of Mechanical Engineering, Indian Institute of Kharagpur, India

Corresponding author: partha.acharya@kgpian.iitkgp.ac.in

---

Demo

Predicted RR

---

Installation

1. Clone the Repository

git clone https://github.com/Nikhil-Rao20/Digital_Twin.git
cd Digital_Twin

2. Create Virtual Environment

python -m venv .venv

# Windows
.venv\Scripts\activate

# Linux/Mac
source .venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Download Datasets

See Dataset/README.md for detailed instructions on downloading:

• BIDMC PPG and Respiration Dataset (PhysioNet)
• CapnoBase Dataset
• COVID-19 Lung CT Data (Kaggle)

---

Usage

Run All Evaluations

Evaluate all respiratory rate extraction methods on the BIDMC dataset:

python scripts/run_all_evaluations.py

Output: Results saved to results/ folder including:

• ppg_window_results.csv - PPG method results
• ecg_edr_butterworth_results.csv - ECG-EDR Butterworth results
• ecg_edr_biquad_results.csv - ECG-EDR Biquad results
• all_results.json - Summary of all methods

---

Run Digital Twin Pipeline

Generate 3D lung breathing animation driven by extracted respiratory rate:

# Full pipeline: extract RR from signal + create animation
python scripts/run_digital_twin.py --lung coronacases_005 --signal bidmc01

# Use constant breathing rate
python scripts/run_digital_twin.py --lung coronacases_005 --rr 15.0

# List available lung datasets
python scripts/run_digital_twin.py --list-lungs

---

Run Prediction Visualization

Visualize lung breathing using predicted respiratory rate from algorithms:

# Using Welch PSD method
python scripts/run_prediction_visualization.py --subject bidmc01 --method welch

# Using ECG-EDR method
python scripts/run_prediction_visualization.py --subject bidmc01 --method ecg_edr

---

Extract RR from a Single Subject

python scripts/extract_sample_rr.py --subject bidmc01

---

Project Structure

Digital_Twin/
├── src/
│   ├── ppg_rr_extraction/    # PPG-based RR extraction (Welch PSD)
│   ├── ecg_edr/              # ECG-derived respiration methods
│   └── lung_twin/            # 3D lung visualization
├── scripts/
│   ├── run_all_evaluations.py
│   ├── run_digital_twin.py
│   └── run_prediction_visualization.py
├── Dataset/                  # Download datasets here
│   └── README.md             # Dataset download instructions
├── assets/                   # Demo videos
├── requirements.txt
└── README.md

---

Methods

Method	MAE (bpm)	Within ±5 bpm
Welch PSD (PPG)	3.20	79.2%
ECG-EDR Butterworth	3.90	68.4%
ECG-EDR Biquad	2.80	84.0%

---

License

This project is for research purposes.