🎭 MoodScope

AI-Powered Real-Time Emotion Detection and Analysis

MoodScope is an intelligent emotion recognition system that integrates face detection, deep learning-based emotion classification, and AI-generated wellness insights.
It can analyze emotions in real time via webcam or from pre-recorded videos / CSV logs, visualize trends, and generate personalized insights using Ollama AI (Llama 3).

---

🚀 Features

🧠 Real-Time Emotion Detection

• Detects faces using MTCNN and classifies emotions with a trained CNN model (emotion_detection_model.h5).
• Displays bounding boxes, emotion labels, and confidence scores directly on live video.
• Supports webcam input or uploaded video files.

📊 Emotion Analytics Dashboard

• Real-time emotion logging to an SQLite database.
• Interactive emotion confidence graph using Matplotlib.
• Exports emotion data to CSV for offline analysis.

🤖 AI-Powered Insights

• Integrates with Ollama (Llama3.2) to generate:
  • Key emotional insights
  • Personalized wellness recommendations
  • Actionable tips for improving emotional wellbeing
• Falls back to rule-based insights if Ollama isn’t running.

🧩 CSV Reflection Mode

• Import CSV emotion logs and visualize trends over time.
• Generates detailed emotional reports and recommendations.

---

🖥️ Tech Stack

Component	Technology
Frontend / UI	PyQt5
Backend / Processing	OpenCV, TensorFlow/Keras
Face Detection	MTCNN
Emotion Model	Custom CNN (emotion_detection_model.h5)
Database	SQLite3
Visualization	Matplotlib
AI Insights	Ollama (Llama3.2 or compatible local model)

---

⚙️ Installation

1️⃣ Clone the Repository

git clone https://github.com/saksham-45/MoodScope.git
cd MoodScope

2️⃣ Create a Virtual Environment

python3 -m venv venv
source venv/bin/activate     # On macOS/Linux
venv\Scripts\activate        # On Windows

3️⃣ Install Dependencies

pip install -r requirements.txt

4️⃣ Download or Train the Emotion Model

Place your trained model in the project root as:

emotion_detection_model.h5


⸻

▶️ Usage

🧍 Real-Time Detection

python main.py

	•	Starts webcam feed
	•	Detects and logs emotions live
	•	Displays emotion trends on a graph

📂 Import and Analyze Emotion Logs

In the startup screen:
	•	Click “📂 Import Emotion Log (CSV)”
	•	Select a previously exported emotion log

⸻

🧠 Optional: Ollama AI Setup (for Insights)

To enable AI-powered insights:
	1.	Install Ollama
	2.	Run Ollama locally:

ollama run llama3.2


	3.	MoodScope will automatically detect and use the local Ollama server at:

http://localhost:11434



⸻

📈 Output Examples

🖼️ Real-Time Detection Window
	•	Bounding boxes with emotion labels and confidence
	•	Dynamic engagement bar and dominant emotion indicator

📊 Analytics Report
	•	Emotion Confidence vs. Time plot
	•	Emotion distribution summary
	•	AI-generated insights and wellness tips

⸻

🧾 Export Options
	•	Export real-time data to CSV (Export CSV button)
	•	Import exported CSVs for reflection and trend analysis

⸻

💡 Future Enhancements
	•	Cloud synchronization for emotion data
	•	Advanced model fine-tuning with FER2013+
	•	Group emotion analytics
	•	API integration for real-world emotion tracking

⸻

🧑‍💻 Author

Saksham Srivastava
📍 Developer & AI Enthusiast

⸻

🪪 License

This project is licensed under the MIT License — feel free to use and modify with attribution.

⸻

🌈 Summary

MoodScope bridges emotion recognition, data visualization, and AI-driven self-awareness.
Whether for research, wellness tracking, or human-computer interaction studies, it offers an all-in-one tool to understand emotions in real time.