Abstract
The integration of artificial intelligence in medical imaging has revolutionized diagnostic capabilities, particularly in cardiology. This study presents an innovative approach to automated cardiac region detection in chest X-ray images using advanced deep learning techniques. We developed and evaluated multiple convolutional neural network architectures, with ResNet-50 emerging as the optimal model for precise cardiac region localization. Utilizing a dataset of 496 high-resolution DICOM images (1024 x 1024), we implemented a comprehensive preprocessing pipeline including intelligent scaling, normalization, and advanced data augmentation techniques. The model achieved remarkable accuracy in predicting bounding box coordinates for cardiac region delineation, demonstrating robust performance across diverse clinical scenarios. Our findings suggest significant potential for improving diagnostic efficiency and accuracy in clinical settings, particularly in resource-constrained environments where expert radiological interpretation may be limited.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call