Hybrid Deep learning based Semi-supervised Model for Medical Imaging

Abstract

One of the most promising fields in medicine is the application of artificial intelligence methods to medical imaging. Though annotating medical images is an expensive operation, most of the recent success in this field has relied heavily on vast amounts of meticulously annotated data. The best that we can tell, the method we present in this research, hybrid approach, is the first to use recent developments in semi-supervised learning (SSL) for medical imaging recognition. Due to the growing complexity of healthcare data, machine learning techniques like Deep Neural Network (DNN) models have become more and more popular. Machine learning (ML) techniques can uncover hidden patterns and other important facts from the massive amount of health data that traditional analytics can’t find in a reasonable length of time. These advancements will have a significant impact on medical imaging technology, medical healthcare data processing, medical illness diagnostics, and general healthcare. We have two goals: First to conduct a survey on DL techniques to medical pictures; and second proposed hybrid DL-based approaches for image classification. This paper mainly proposed hybrid methodology for medical imaging which can improve semi supervised model.