🧠 MoodGuard · AI Mental Health Monitor

Real-time emotion detection powered by deep learning and computer vision — analyze facial expressions through photos or live webcam using AI.

---

📌 Overview

MoodGuard is an AI-powered emotion recognition and mental health monitoring web application that uses Convolutional Neural Networks (CNN) and computer vision to detect facial emotions in real time.

The system can analyze uploaded images, captured photos, and live webcam feeds to predict emotional states such as happiness, sadness, anger, fear, surprise, disgust, and neutrality.

It is designed with a privacy-first local AI approach, meaning no facial data is sent to any external server.

Metric	Value
🎯 Model Accuracy	66.6%
🖼️ Training Images	35,000+
😄 Emotion Classes	7
🧠 Model Type	CNN
📦 Dataset	FER2013
🔒 Privacy	100% Local Processing

---

✨ Features

📸 Photo Analysis

Upload facial images and instantly detect emotional state with AI-powered prediction.

🎥 Live Webcam Detection

Real-time emotion detection through webcam feed.

📊 Emotion Prediction Dashboard

Clean and interactive dashboard UI built with HTML, CSS, and JavaScript.

🕒 Detection History

Can track and display previous analysis sessions.

🔒 Privacy First

All processing happens locally. No cloud upload. No external APIs.

⚡ Fast AI Inference

TensorFlow model loaded directly in Flask for fast prediction.

🧠 Deep Learning Model

CNN trained on FER2013 facial expression dataset.

---

🛠️ Technology Stack

Technology	Role
🐍 Python 3.10	Core language
🌐 Flask	Backend web framework
🤖 TensorFlow / Keras	CNN model loading & prediction
📷 OpenCV	Face detection & image preprocessing
🔢 NumPy	Numerical computation
🖼️ Pillow	Image handling
🎨 HTML / CSS / JS	Frontend dashboard
📦 H5Py	Loading .h5 trained model

---

🚀 Getting Started

Prerequisites

• Python 3.10+
• Webcam (for video detection)
• moodguard_model.h5

---

Installation

1. Clone Repository

git clone https://github.com/yourusername/MoodGuard.git
cd MoodGuard

2. Create Virtual Environment

python -m venv venv

Windows

venv\Scripts\activate

Mac/Linux

source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

4. Add Model File

Place:

moodguard_model.h5

inside project root directory.

---

▶ Run Project

python app.py

Open browser:

http://127.0.0.1:5000

---

📦 Requirements

Flask==3.0.3
tensorflow==2.10.0
opencv-python==4.10.0.84
numpy==1.23.5
Pillow==10.4.0
h5py==3.11.0
gunicorn==22.0.0

Install:

pip install -r requirements.txt

---

🔬 Detection Pipeline

Input Image / Webcam
        │
        ▼
┌────────────────────────┐
│ 1. Image Capture       │
│ Photo Upload / Webcam  │
└────────────┬───────────┘
             │
             ▼
┌────────────────────────┐
│ 2. Face Detection      │
│ OpenCV Haar Cascade    │
└────────────┬───────────┘
             │
             ▼
┌────────────────────────┐
│ 3. Preprocessing       │
│ Resize + Normalize     │
└────────────┬───────────┘
             │
             ▼
┌────────────────────────┐
│ 4. CNN Prediction      │
│ TensorFlow / Keras     │
└────────────┬───────────┘
             │
             ▼
┌────────────────────────┐
│ 5. Emotion Output      │
│ UI + Prediction Result │
└────────────────────────┘

---

😄 Supported Emotions

MoodGuard detects 7 emotions:

• Happy 😊
• Sad 😢
• Angry 😠
• Fear 😨
• Surprise 😲
• Neutral 😐
• Disgust 😖

---

🧠 Model Details

Property	Value
Model	CNN
Framework	TensorFlow / Keras
Dataset	FER2013
Classes	7 Emotions
Training Images	35K+
Format	.h5
Input	Facial Expressions

---

📁 Project Structure

MoodGuard/
│
├── app.py
├── detect_emotion.py
├── prepare_dataset.py
├── train_model.py
├── moodguard_model.h5
├── requirements.txt
├── README.md
├── .gitignore
│
├── templates/
│   └── index.html
│
└── static/
    ├── style.css
    └── script.js

---

🖥️ App Sections

Section	Description
📸 Photo Analysis	Upload image → detect emotion
🎥 Video Detection	Real-time webcam prediction
📜 History	Previous detections
ℹ️ About	System overview

---

⚠️ Troubleshooting

Model file not found

Place moodguard_model.h5 in root directory.

Camera not opening

Check OS/browser camera permissions.

No face detected

Use a clear front-facing image.

Slow prediction

Close other heavy applications.

TensorFlow install issue

Use Python 3.10 (recommended).

---

🗺️ Roadmap

• Mood trend analysis
• PDF reports
• Session analytics
• Database support
• Dark mode
• Login system
• Voice emotion detection
• AI chatbot integration
• Docker deployment

---

👩‍💻 About Creator

Hamna Munir
AI/ML Engineer • Software Engineering Student

Built MoodGuard as an end-to-end AI project combining:

• Deep Learning
• CNN Architecture
• Computer Vision
• Flask Deployment
• Frontend + Backend Integration
• Real-time AI Prediction

This project demonstrates a complete machine learning workflow from preprocessing → model training → deployment.

---

📄 License

Licensed under MIT License.

---

🙏 Acknowledgements

• FER2013 Dataset
• TensorFlow
• OpenCV
• Flask
• Keras
• Open-source AI/ML community

---