International Journal of Science, Engineering and TechnologyAuthors: Mrs. D.R. Nanda Devi, Gnaneshwari Sarala, K. Yasaswitha, A. Likhitha
Abstract: In the context of the current healthcare environment, the accurate diagnosis of a patient's health condition is only possible by the convergence of the self-reported symptoms by the patient and the analysis of medical images such as a chest X-ray, CT scans, or MRI scans. However, the current state-of-the-art models only provide the analysis of self-reported symptoms by the patient using a natural language-based system, which responds to the queries made by the patient (for instance, using transformer-based models such as BERT or GPT), or the analysis of medical images using a vision-based system (for instance, using ResNet or EfficientNet models). Therefore, the current fragmented approach not only affects the potential of accurate diagnosis by using AI-based models but also affects the comprehensive analysis carried out by the diagnostic professionals in the context of self-reported symptoms as well as medical images. A comprehensive analysis system using medical images as well as natural languages is not available in the current context; hence, the accessibility of healthcare has become a significant concern.
