BlindGate - An iOS App for Object Recognition and Accessibility

BlindGate is an iOS application designed to assist blind and visually impaired individuals by identifying objects in their surroundings and providing voice-based feedback. This app leverages YOLOv5 for real-time object detection and accessibility features to enhance independence and situational awareness.

---

Features

• Object Recognition: Uses the device camera and YOLOv5 to detect and identify objects in real-time.
• Voice Feedback: Provides auditory descriptions of identified objects using text-to-speech.
• Accessibility: Fully compatible with VoiceOver and provides haptic feedback for better interaction.
• User-Friendly Interface: Minimalistic design with gesture-based controls for easy navigation.

---

Object Detection in Action

---

Technology Stack

• Language: Swift
• Frameworks and Tools:
  • YOLOv5: For object detection using a pre-trained model exported to CoreML.
  • Vision Framework: To integrate YOLOv5 for on-device inference.
  • AVFoundation: For text-to-speech feedback.
  • SwiftUI: For creating an accessible and intuitive user interface.

---

How It Works

1. Object Detection:

   • YOLOv5 is used for real-time object recognition.
   • The YOLOv5 PyTorch model is converted to CoreML format to run natively on iOS devices.

2. Voice Feedback:

   • Identified objects are described to the user through iOS's text-to-speech functionality.

3. Accessibility:

   • Full VoiceOver support is provided for visually impaired users.
   • Haptic feedback is used to notify users of successful detections or errors.

---

Getting Started

Prerequisites

• macOS with Xcode installed.
• An iOS device with a camera (the app requires camera access).
• YOLOv5 CoreML model file (.mlmodel).

Installation

1. Clone this repository:

   git clone https://github.com/sandy787/BlindGate.git