Hiatal Hernia Detection Using Novel End-to-End Deep Learning Model with Explainable AI

Hiatal hernia is a condition where the stomach or other internal organs push through the diaphragm into the chest. It is still relatively unexplored area with scanty research making it crucial to enhance the efficiency of detecting this condition by automating the task and providing preliminary diagnosis. Hence, this paper proposes a novel approach to detect hiatal hernias using Deep Learning and Python. This model seeks to increase the accuracy of hiatal hernia detection using explainable AI technology. Compared to existing pretrained models, creating a custom model reduces training time. The model also facilitates integrated preprocessing, which enhances image contrast and provides more information for the learning procedure. This paper presents an innovative methodology for the detection of Hiatal Hernia leveraging Deep Learning and Explainable AI techniques. We achieved a robust classification of positive and negative cases of Hiatal Hernia using ChestX-rays, attaining a notably high level of accuracy. To provide transparency and interpretability in our model's predictions, we employed LIME, an Explainable AI method. This approach facilitated a deeper understanding of the influential features that contributed to the prediction outcomes. These findings collectively demonstrate the efficacy of our approach, offering a promising avenue for more accurate and expedited Hiatal Hernia detection, while also providing valuable insights into the diagnostic process.