PERFORMANCE COMPARISON OF MACHINE LEARNING MODELS IN MEDICAL IMAGE RECOGNITION

This paper presents a comparison of the ML models for medical image recognition tasks. Conventional techniques e.g. SVM and Random Forest and also recent deep learning structures e.g. CNNs are analysed. Model accuracy, precision, recall, F1-score, inference time and performance are measured, and the models are verified, and benchmark medical imaging datasets are validated to ensure accuracy of the image. Both approaches have their advantages and disadvantages and the presented methods provide an insight into the applied use of each method in automated medical diagnosis.