# 📊 NeuroEcho: Multimodal Emotion Consistency Tracking for Telehealth **NeuroEcho** is an AI-powered system designed to track, analyze and visualize emotional signals from patients during virtual sessions in real time. It processes live webcam and microphone input to assess emotions from facial expressions, vocal tone, and spoken content. Its unique **Emotional Consistency Index (ECI)** highlights discrepancies between what people say and what they feel — offering clinicians deeper emotional insight. This guide will help you set up **NeuroEcho** on any **Windows laptop with Python 3.10** — whether you have an NVIDIA GPU or not. --- ## 🖥️ System Requirements * **Windows 10 / 11** * **Python 3.10.x** (must be installed) * **Git** (for cloning the repository) * **Webcam & Microphone** * **Internet connection** (for downloading models initially) * **Optional:** NVIDIA GPU + CUDA drivers (if available) --- ## 📦 Step-by-Step Installation (Windows) ### 📥 1️⃣ Clone the Repository Open **Command Prompt** and run: ```bash git clone https://github.com/deeptesh-rout/NeuroEcho cd NeuroEcho ``` ### 🐍 2️⃣ Create a Python 3.10 Virtual Environment ```bash python -m venv .venv ``` ### ⚙️ 3️⃣ Activate the Virtual Environment ```bash .venv\Scripts\activate ``` You should now see: ``` (.venv) C:\Users\YourName\NeuroEcho> ``` ### 🔍 4️⃣ Confirm Python Version ```bash python --version ``` ✅ It must say **Python 3.10.x** ### 📜 5️⃣ Install Python Dependencies ```bash pip install -r requirements.txt ``` --- ## ⚡ Install PyTorch (Choose based on your hardware) ### 🖥️ A) If you **DO NOT HAVE a GPU (CPU-only)** Run: ```bash pip install torch==2.7.0 torchaudio==2.7.0 --index-url https://download.pytorch.org/whl/cpu ``` ### 🖥️ B) If you **HAVE an NVIDIA GPU** (with CUDA drivers installed) * First, check your CUDA version: * Open **Command Prompt** and run: ```bash nvidia-smi ``` * Note the CUDA version listed (e.g., 12.8) * Then run the corresponding command: **Example for CUDA 12.8:** ```bash pip install torch==2.7.0 torchaudio==2.7.0 --index-url https://download.pytorch.org/whl/cu128 ``` **Replace `cu128` with your actual CUDA version number if different.** --- ## 🎥 6️⃣ Install FFmpeg (Required for Audio/Video Processing) * Download the Windows build from: [https://www.gyan.dev/ffmpeg/builds/](https://www.gyan.dev/ffmpeg/builds/) * Extract it to `C:\ffmpeg` * Add `C:\ffmpeg\bin` to your **System PATH** * Test it by running: ```bash ffmpeg -version ``` ✅ You should see version info. --- ## 🚀 Running NeuroEcho With your virtual environment activated: ```bash python app/run_affectlink.py ``` ✅ The program will: * Start audio/video processing threads * Perform Whisper transcription, Speech Emotion Recognition (SER), and Text Emotion detection * Open a **Streamlit dashboard** at `http://localhost:8501` If it doesn't auto-open — open your browser and navigate to: ``` http://localhost:8501 ``` --- ## 🛑 Stopping the Application In the Command Prompt where it's running, press: ``` Ctrl + C ``` This will stop all services gracefully. --- ## 📦 Optional: Disable Pylance Type Warnings (VS Code Users) In **VS Code** settings: ``` File → Preferences → Settings → Search 'typeCheckingMode' → set to 'basic' or 'off' ``` --- ## ✅ Recap Summary | Step | Command / Action | | :--------------------------------- | :------------------------------------- | | Clone project | `git clone ...` | | Create venv | `python -m venv .venv` | | Activate venv | `.venv\Scripts\activate` | | Confirm Python version | `python --version` → 3.10.x | | Install requirements | `pip install -r requirements.txt` | | Install CPU PyTorch | `pip install torch==2.7.0 ... --cpu` | | Install GPU PyTorch (if available) | `pip install torch==2.7.0 ... --cuXXX` | | Install FFmpeg | Download + add to PATH | | Run the program | `python app/run_affectlink.py` | | Open dashboard | `http://localhost:8501` | | Stop program | `Ctrl + C` | --- ## 📜 License This project is licensed under the MIT License. --- ## 💡 About NeuroEcho **NeuroEcho** bridges the empathy gap in telehealth with real-time multimodal emotion detection. It processes facial cues, voice signals, and spoken text to calculate an Emotional Consistency Index (ECI) — helping clinicians better understand their patients' unspoken feelings. Built with ❤️ by Deeptesh. --- 🎉 You’re all set. Enjoy NeuroEcho! # NeuroEcho