Automated bone age assessment from hand radiographs is a clinically valuable yet challenging task due to the variability in radiograph quality, anatomical presentation, and the subjectivity inherent in manual evaluation methods. In this work, we present an end-to-end deep learning pipeline for the preprocessing, segmentation, and estimation of bone age from hand X-ray images. The proposed methodology leverages the RSNA Pediatric Bone Age Challenge 2017 dataset [5]. It integrates advanced preprocessing techniques to standardize radiograph quality, followed by a U-Net-based model to extract key hand regions. Separate regression models are trained on these segmented regions, and their outputs are combined in a fusion model using ResNet50, incorporating patient gender as an additional input. Experimental results demonstrate that the fused model achieves a mean absolute error (MAE) of 15.36 months on the validation set, outperforming models trained on individual regions though not yet state-of-the-art. The pipeline also demonstrates computational efficiency, supporting its potential integration into clinical workflows. Qualitative analysis of saliency maps confirms that the model focuses on clinically relevant areas, reinforcing the approach’s interpretability. Overall, this end-to-end pipeline provides a robust and objective framework for automated bone age assessment from hand radiographs.