MoodMind

What it does

MoodMind is an emotion detection system integrated with personalized recommendations based on the detected emotions. The system uses `cv2` (OpenCV) for capturing video frames and `DeepFace` for emotion analysis.

Based on the detected emotion, the Gemini API is called to generate a recommendation. The API call requests a short, five-word recommendation tailored to the specific emotion. The recommendations are dynamically generated for each detected emotion: happy, sad, angry, neutral, surprise, fear, and disgust.

The captured video frame is displayed with overlay text showing the user's name, age, detected emotion, and the corresponding recommendation from the Gemini API. The video feed updates in real-time, continuously analyzing and providing recommendations until the user decides to quit the session by closing the video feed.

The Gemini model (`gemini-1.5-flash`) is initialized to handle the content generation. For each detected emotion, the model's `generate_content` method is called with a prompt to generate a short recommendation. The generated text is then displayed on the video feed.

Additionally, the system stores user data, including name, age, detected emotion, and recommendation, in Firebase Firestore. The video feed updates every 10 seconds, capturing the current frame, analyzing the emotion, generating a recommendation, and storing the data in Firestore.