Abstract
In this work, InceptionV3 is applied to identify cardiomegaly from chest X -ray images through a deep learning technique. The dataset, sourced from Kaggle, contains 5,552 balanced images across two classes: The presence of cardiomegaly, coded as "True" and the absence of cardiomegaly coded as "False". Some pre-processing steps which include image resizing, normalization and augmentation improved the performance of the proposed model. The model was trained in 10 epochs with a batch size of 64 and produced an overall accuracy of 86%. The evaluation criteria such as precision, recall, F1-score, confusion matrix and classification report proved high and fair results on both classes. The results show a slight overfitting problem, which needs to be address in the future work by using approaches such as regularization and dropout. Future work includes using real and bigger datasets, combining different types of data, and using eXplainable AI for increased model interpretability. This work establishes the possibility of using deep learning for automating the identification of cardiomegaly, which can be an important direction for the development of efficient and accurate diagnostic tools in medical imaging.