A Deep-Learning Model Based on Channel and Spatial Attention for Cervical Fractures Prediction Using CT Images

Traumatic cervical spine vertebrae injuries are dangerous and require accurate and quick diagnosis for treatment. The majority of cervical spinal cord injuries, 56%, are cervical spine fractures, which increase trauma patient morbidity and death. Given the importance of this healthcare environment medical issue, our research suggests the CNN with Channel and Spatial Attention-based solution. The neural network architecture of the proposed model is designed to identify cervical spine fractures in CT images. To enhance cervical fracture predictions, this model uses channel and spatial attention processes in a Convolutional neural network to focus on key features and spatial areas in pictures. It had 99.50% accuracy, 99.50% precision, and 97.25% recall on a large dataset of 4200 CT scans from Kaggle. A low loss function of 0.075% highlights the model's amazing learning ability. The model method, which emphasizes attention mechanisms, advances medical imaging and fracture detection, improving cervical spine injury diagnosis.