Classification Toolkit for Medical Imaging Researchers

This project develops a comprehensive toolbox for applying machine learning algorithms to medical imaging, facilitating advanced analysis and predictions across various medical imaging datasets.

Machine learning plays a crucial role in the field of digital image processing, particularly within the medical domain. This project introduces a toolbox designed to assist researchers in applying different machine learning algorithms on medical images, enabling predictions and analysis across diverse medical image datasets.

With the absence of a GUI tool for applying machine learning techniques to medical images, this project aims to fill this gap by providing a toolkit for preprocessing, feature extraction, training, testing classifiers, and achieving results. It supports loading various datasets and applying multiple technique combinations to medical imaging datasets.

The toolkit encompasses several key functionalities:

• Data Loading: Allows users to select datasets stored on their devices, providing insights into the total number of instances and classes within the dataset.
• Preprocessing: Implements various filtering and segmentation techniques, such as Gaussian, Sobel, and Canny filtering, along with K-Means and Otsu segmentation.
• Feature Extraction: Extracts both handcrafted features (e.g., HOG and LBP features) and deep neural network features (e.g., AlexNet and ResNet features).
• Training and Testing: Facilitates training of models selected by the user and evaluates their accuracy.

Conclusion

The Automated Classification Toolkit for Medical Imaging Researchers stands as a significant advancement for researchers in the medical imaging field, providing a robust platform for applying machine learning algorithms to medical images.

Check out the presentation for more details: