📸 Better Boyfriend

The AI-powered camera assistant that saves relationships, one photo at a time.

📖 Overview

Better Boyfriend is a computer vision web application designed to solve a universal problem: bad photos taken by reluctant partners.

Unlike standard camera apps that just provide a grid, this application uses Real-Time Edge AI to actively "coach" the photographer. It tracks the subject's skeleton in 30fps, compares it against professional pose templates, and provides granular, real-time feedback (e.g., "Tilt Down," "Step Back," "Perfect") to ensure the perfect shot.

🚀 Live Demo

Try the latest build on your mobile device: https://betterboyfriend.netlify.app

Note: Please allow camera permissions when prompted. The AI runs entirely locally on your device for privacy.

---

✨ Features

• Real-Time Skeletal Tracking: Detects 33 distinct body landmarks (nose, wrists, hips, ankles) instantly using the front or back camera.
• "Traffic Light" Feedback System:
  • 🔴 Red: Subject not found.
  • 🟡 Yellow: Subject detected but pose is incorrect (e.g., "Too close," "Headroom issue").
  • 🟢 Green: Perfect composition.
• Smart Auto-Capture: Automatically triggers a 3-second countdown when the perfect angle is held.
• Guided Poses:
  • The Headshot: Centers facial features.
  • Tourist Hand: Ensures hand placement interacts correctly with landmarks.
  • Full Body (OOTD): Validates head-to-toe visibility.
  • Power Pose: Detects confident body language.

---

🛠️ Technical Architecture

This prototype is built as a Single-Page Application (SPA) designed for high-performance mobile usage without backend latency.

The Stack

• Frontend: HTML5, Vanilla JavaScript (ES6+).
• Styling: Tailwind CSS (via CDN) for rapid, glassmorphic iOS-style UI.
• Computer Vision: Google MediaPipe (Pose Landmarker Task).
• Rendering: HTML5 Canvas API (for skeletal overlays) + SVG (for guide silhouettes).
• Hosting: Netlify (Static Edge Hosting).

---

💻 Local Installation

To run this project locally on your machine:

1. Clone the repository:

   git clone [https://github.com/your-username/better-boyfriend.git](https://github.com/your-username/better-boyfriend.git)

2. Navigate to the directory:

   cd better-boyfriend

3. Serve the file: Because this uses the Camera API, it requires a secure context (HTTPS or localhost). You cannot just double-click the HTML file.

   Python 3:

   python3 -m http.server

   VS Code: Right-click index.html and select "Open with Live Server".

4. Open in Browser: Go to http://localhost:8000.

---

🔮 Roadmap: Path to App Store

Moving this from a Web Prototype to a fully scalable iOS/Android application involves the following infrastructure evolution:

Phase 1: Native Wrapper

• Tech: React Native or Flutter.
• Goal: Wrap existing logic to access native camera controls (focus, exposure) and save directly to Camera Roll.

Phase 2: Cloud & Backend

• Auth: Firebase Auth for user accounts.
• Database: Firestore to store custom pose vector data.
• Storage: AWS S3 for backing up high-res images.

Phase 3: Advanced AI

• Training: Train custom TensorFlow Lite models on "Instagram-worthy" datasets.
• Generative AI: Implement Stable Diffusion API for real-time lighting correction.

---

🤝 Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

Authors & Presentation

Nitheesh, Sushant, Peeyush Kumar, Alex zhang, Anvesh

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.