TFLite2-TurkishCoinDetection

TFLite2-TurkishCoinDetection is an AI model designed to detect and classify Turkish Lira coins using TensorFlow Lite. The model is lightweight and optimized for deployment on resource-constrained devices such as mobile and IoT systems.

Features

• Turkish Lira Coin Detection: Accurately classifies coins like 1 TL and 50 kuruş.
• TensorFlow Lite Support: Optimized for mobile and embedded systems.
• Simplified Training Workflow: Supports data labeling using tools like LabelImg and Roboflow.

Technical Details

• Model: SSD Mobilenet V2 FPNLite 320x320
• Training: Built using TensorFlow Object Detection API and converted to TFLite format.
• Performance:
  • Accuracy: Over 90%
  • mAP (mean Average Precision): 85%

Use Cases

• Mobile applications to assist visually impaired individuals in recognizing coins.
• Real-time object detection systems for Turkish coins.

Installation and Usage

Requirements

• Python 3.7+
• TensorFlow
• OpenCV

Train the Model

Run the coin_model.ipynb file in Jupyter Notebook to start the training process.

Example Usage

Integrate the trained model into a mobile application to perform coin recognition using live camera feeds.

Future Work

• Enhance model performance with a more extensive and diverse dataset.
• Optimize for real-time recognition capabilities.
• Expand the model to recognize coins from different countries.

See you next project:)