SaifGotrane/Tunisian-Coin-Detector

# Tunisian Coin Detector This Android application utilizes the YOLOv8 model to perform real-time detection and recognition of Tunisian coins. The app uses the device's camera to capture video, identify different coin denominations, and calculate the total sum of the detected coins. ## Features * **Real-time Coin Detection:** Uses the camera feed to detect Tunisian coins in real-time. * **Coin Recognition:** Identifies the denomination of each detected coin. * **Total Sum Calculation:** Calculates and displays the total value of the detected coins. * **TensorFlow Lite Integration:** Employs a YOLOv8 model converted to TensorFlow Lite for efficient on-device inference. * **CameraX:** Built with Android CameraX for an efficient and easy-to-use camera implementation. ## How it Works 1. **Camera Input:** The application uses the device's camera to capture a live video stream. 2. **Image Processing:** Each frame from the camera feed is pre-processed and converted into a format suitable for the YOLOv8 model. 3. **Object Detection:** The pre-processed image is fed into the TensorFlow Lite YOLOv8 model, which detects and localizes coins within the frame. 4. **Result Display:** The detected coins are highlighted with bounding boxes, and the denomination of each coin is displayed. The total sum of all detected coins is also calculated and shown on the screen. ## Project Structure The project is structured as a standard Android application with the following key components: * `MainActivity.kt`: The main entry point of the application, responsible for managing the camera, UI, and user interactions. * `Detector.kt`: A core class that handles the loading of the YOLOv8 model, pre-processing of the input images, running the inference, and post-processing the detection results. * `BoundingBox.kt`: A data class to represent the bounding box information for each detected object. * `Constants.kt`: A file to store constants used throughout the application, s …