Skip to content

This Android app uses the YOLO model for real-time object detection, built in Kotlin. It accurately identifies and classifies objects through the camera, providing fast and efficient results directly on your device.

Notifications You must be signed in to change notification settings

Alenaak/Detectify

Repository files navigation

Detectify

This Android app allows users to detect objects in real-time using both the front and back cameras with the YOLO (You Only Look Once) model. The app provides a smooth interface for switching cameras and visualizing detected objects with bounding boxes and labels. It is optimized for fast, accurate detection on mobile devices.


Features

  1. Dual Camera Support: Easily switch between the front and back cameras for object detection.
  2. Real-Time Object Detection: Utilizes the YOLO model to detect multiple objects in real-time with high accuracy.
  3. Bounding Boxes and Labels: Detected objects are highlighted with bounding boxes and class labels for better visualization.
  4. Efficient Processing: Powered by TensorFlow Lite for smooth and efficient model inference on Android devices.
  5. User-Friendly Interface: Designed with Jetpack Compose for a clean, responsive, and intuitive UI.

Technical Implementation

YOLO Model Integration

  • The app integrates the YOLOv5 model converted to TensorFlow Lite for fast, on-device inference.
  • Models are optimized for mobile devices to balance accuracy and performance.

Object Visualization

  • Detected objects are overlaid with dynamic bounding boxes and labels.
  • The UI updates instantly with every frame, ensuring smooth object tracking.

Performance Optimization

  • Efficient use of background threads for model processing to keep the UI responsive.
  • Uses TensorFlow Lite’s GPU delegate for accelerated performance (if available).

User Experience

  1. Launch the App → Grant Camera Permission → Real-time Object Detection.
  2. Use the Switch Camera button to toggle between front and back cameras.
  3. View detected objects with live bounding boxes and labels.

Application Interface


Error Handling

  • Camera Access Denied: Prompts the user to grant camera permissions if denied.

Conclusion

The Detectify App is a powerful and efficient app for real-time object detection using both the front and back cameras. By combining the YOLO model with TensorFlow Lite and Camera, the app offers accurate detection with a user-friendly interface, making it perfect for everyday use and practical applications in object recognition.

About

This Android app uses the YOLO model for real-time object detection, built in Kotlin. It accurately identifies and classifies objects through the camera, providing fast and efficient results directly on your device.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages