The AI Classifier App is a graphical interface application designed to identify specific body parts in medical images. It employs deep learning techniques to train a robust model that classifies images into predefined categories. This project aims to enhance medical imaging analysis and aid in diagnostic workflows.
-
GUI Interface: User-friendly interface for training, testing, and managing the model.
-
Deep Learning Integration: Utilizes PyTorch and ResNet18 for model training and inference.
-
Image Preprocessing: Includes advanced preprocessing pipelines for training and testing.
-
Customizable Dataset: Supports dynamic addition of labeled images for training.
-
Real-time Feedback: Displays training progress and prediction results in real time.
-
Logging and Persistence: Saves training data and model checkpoints to disk for future use.
-
Python 3.8 or higher
-
A CUDA-compatible GPU is recommended for faster training.
-
Required libraries (listed in requirement.txt )
The application logs key information including:
-
Training progress and accuracy.
-
Prediction results with probabilities.
-
Errors encountered during image loading or model training.
-
Logs are displayed in the GUI and saved in structured formats for analysis.
- The application uses a dataset divided into labeled categories, with each category corresponding to a body part. Data is dynamically managed through the GUI and saved in JSON format for persistence.
- We welcome contributions! To contribute: