Evaluation Study for Celiac Disease Diagnosing by using Deep Learning Techniques

Celiac disease (CD) is the autoimmune reaction that occurs as a result of ingestion of gluten, which results to mucosal injury. Proper screening is especially important here, though the existing diagnostic techniques are frequently expensive, as well as the time-consuming. This review seeks to uncover the future of deep learning techniques in changing the diagnosis of CD through imagery of biopsies. we explore Single Layer ABC-DL architectures, CNN in CD diagnosis, Transfer Learning and Multiple Instance Learning (MIL). In this review, these methods are described with regard to their current capabilities to improve diagnostic accuracy and speed over conventional approaches compared to those used in the construction of the models. Additionally, we describe the prospects and flaws seen in applying DL to medical image analysis and the critical issues of increased requirements for big labeled datasets, data quality, and the interpretability of the model. Last but not least, we discuss the possible future research avenues of the DL applications for CD diagnosis like multi-modal approach and better interpretability mechanisms. This review enable the author to identify the current status of developing deep learning techniques for CD diagnosis and further research direction in this promising field.