SenseRoute is a smart, assistive virtual assistant designed for the visually impaired. It provides real-time environmental awareness, object recognition, voice interaction, and emergency alerting — empowering users to navigate the world independently and safely using AI and speech technologies.

• 🎯 Object Detection
  Detects and recognizes real-world objects using YOLOv8 and OpenCV.

• 🗣️ Text-to-Speech (TTS)
  Converts scene and object information into speech for auditory feedback.

• 📧 Emergency Email Alerts
  Sends alerts with object descriptions and timestamps to a guardian via Gmail API.

• 🔊 Voice Command Integration
  Responds to custom voice commands like "describe environment", "detect object" etc.

• 📰 Location-Based News Updates
  Reads out the latest news headlines relevant to the user’s city using NewsAPI.

• 📄 OCR (Optical Character Recognition)
  Reads and speaks text from images using Tesseract OCR.

• 🎮 Maze Game
  A fun, brain-engaging game to improve spatial reasoning and navigation.

• 📷 Scene Description
  Captures and summarizes the environment around the user.

• 🔒 Secure & Private
  API keys and tokens are excluded from the repository for safety.

• Frontend: React.js, Tailwind CSS, ShadCN UI, Framer Motion
• Backend: Python (Flask), Node.js (optional)
• AI & Vision: YOLOv8, OpenCV, Tesseract OCR
• Voice Tools: gTTS / pyttsx3, SpeechRecognition
• APIs Used:
  • Gmail API – For sending alerts
  • NewsAPI – For local headlines
  • OpenWeather API (optional) – For weather updates
• Other Tools:
  • GitHub Actions (CI/CD)
  • Google OAuth (secure access)
  • Git Filter Repo (secret removal)

SenseRoute/
├── frontend/                # React UI
├── assistant-backend/      # Python Flask backend
│   ├── object_detection/   # YOLOv8 integration
│   ├── ocr_module/         # Tesseract OCR logic
│   ├── email_service/      # Gmail API handling
│   └── news_module/        # Location-based news reader
├── .spyder-py3/            # Local config (ignored in Git)
├── credentials.json        # (ignored) Google OAuth credentials
├── token.json              # (ignored) Gmail API token
├── README.md               # Project documentation
└── .gitignore              # Git ignore rules

This blind-accessible maze game empowers visually impaired users to explore and solve mazes using audio feedback and keyboard navigation.

• ✅ Spatial Audio Guidance

  • As the player (👧) gets closer to the goal (🏠), the volume increases.
  • Bumping into walls triggers sound effects from the direction of the wall, aiding spatial understanding.

• ✅ Keyboard-Based Navigation

  • Press Shift for Previous Level
  • Movement press (↑ ↓ ← →)
  • Press Enter to New Maze
  • No mouse or screen interaction required

• ✅ Maze Generation

  • Mazes are generated using a depth-first search algorithm
  • Ensures random, solvable mazes each time

• ✅ Level Progression

  • 4 levels of increasing difficulty
  • Complete 3 mazes per level to unlock the next (12 total)

• ✅ Replayability

  • Regenerate mazes anytime
  • Replay previous levels to improve skills

Help visually impaired users reach the destination (🏠) using audio cues and keyboard controls.
This interactive game boosts confidence, spatial reasoning, and provides a fun, skill-building experience.