Multi-Factor Classification Using Deep Learning for X-Ray Image Classification

Medical imaging is a pivotal tool in modern diagnostics, offering detailed visualization of internal body structures. Skeletal radiographs (X-rays) are essential for assessing bone conditions and anomalies. This research aims to enhance diagnostic precision and efficiency by leveraging deep learning models to predict patients' age and gender from chest radiographs. We aim to evaluate the performance of pre-trained models (VGG-16, MobileNet, EfficientNetB0) against a custom ResNet-50 model designed for skeletal radiograph analysis. A dataset of chest X-ray images was collected and preprocessed by converting to grayscale, resizing, and normalizing to ensure consistency and quality. The study employs performance metrics, including mean absolute error (MAE) and accuracy, to comprehensively evaluate the models. The VGG-16 model demonstrated the highest performance, achieving an age prediction MAE of 9.03 and a gender classification accuracy of 83.96%. This research underscores the critical roles of dataset quality and model architecture in improving predictive accuracy, providing valuable insights for developing more effective deep learning models in medical imaging.