Leveraging Spatial and Temporal Features using CNN-LSTM for Improved Bone Fracture Classification from X-ray Images

X-ray imaging remains the primary diagnostic tool for identifying bone fractures. Accurate classification of bone fractures from X-ray images is a critical task in the field of orthopedic diagnosis and treatment. However, this task poses several challenges due to the complex and variable nature of fracture patterns, as well as the need to consider the temporal progression of fracture healing. Traditional convolutional neural network (CNN) architectures, while effective in extracting spatial features from X-ray images, may not be sufficient to capture the dynamic and sequential aspects of fracture healing. In this work, we propose an improved Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) architecture to address the limitations of CNN-only models in bone fracture classification. The proposed method aims to leverage the strengths of both CNNs and LSTMs to enhance the classification performance. The CNN component of the network is responsible for extracting relevant visual features from the X-ray images, while the LSTM layers are used to model the temporal relationships between these features, which are crucial for understanding the progression of fracture healing. Experimental results on a dataset of X-ray images of bone fractures demonstrate the superior performance of the CNN-LSTM architecture compared to traditional CNN-based approaches with 97.33%