TriageTech: AI-Based Medical Triage System

TriageTech is a smart patient triaging tool designed to prioritize patient treatment using real-time sensor data and AI-driven classification. This project uses a machine learning model to categorize patients into emergency classes based on vitals like temperature, SpO₂, blood pressure, heart rate, and age.

Project Overview

The system includes:

• A serial interface to collect real-time vitals from medical sensors.
• An MLP-based neural network model to predict triage priority.
• A web-based medical form for user input.
• An automated Excel-based priority list generator.

File Structure

File	Description
ANN.py	Main AI pipeline: trains an MLP classifier on vitals data, predicts patient class, and saves a prioritized list.
pyserial_final.py	Collects sensor data via serial port and appends it to sensor_data.xlsx.
TriageForm2.html	Front-end form to collect patient metadata and symptoms.
sensor_data_training.xlsx	Training dataset for model training.
sensor_data.xlsx	Live input from sensors during testing.
patient_priority_list.xlsx	Output file with predicted triage priorities.

Model Summary

• Type: Multi-layer Perceptron (MLP)
• Framework: TensorFlow / Keras
• Input Features: Temperature, SpO₂, BP, HR, Age
• Classes:
  • Class A: Emergency (0–5 min)
  • Class B: Urgent (10–20 min)
  • Class C: Non-Urgent (30+ min)
• Output: Severity class, triage priority, and estimated wait time

How to Run

1. Prerequisites

pip install pandas numpy openpyxl matplotlib scikit-learn keras tensorflow pyserial

2. Collect Sensor Data

Run this script to log real-time data from a serial device (e.g., Arduino with sensors):

python pyserial_final.py

Ensure the device outputs data in the format:

Temperature=36.5
SpO2=98
Heart Rate=72

3. Train and Predict Using ANN

Train the neural network and generate the triage priority list:

python ANN.py

This will output patient_priority_list.xlsx with predictions.

4. Web Form Usage (Optional)

Open TriageForm2.html in a browser to input patient details manually. Submissions will be sent to a connected backend or script (Google Apps Script in this case).

Output Example

Temp	SpO₂	BP	HR	Age	Severity	Priority	Wait Time
39.0	85	190	105	72	Emergency (A)	1	0-5 min
37.0	92	160	90	55	Urgent (B)	2	10-20 min
36.5	98	120	72	24	Non-Urgent (C)	3	30+ min

Highlights

• Real-time triage support using AI
• Dynamic input via serial port
• GUI-ready HTML form
• Excel integration for real-world hospital use

Tech Stack

• Python
• TensorFlow/Keras
• Pandas / NumPy
• pySerial
• HTML + Bulma CSS
• Google Apps Script (form backend)