Evaluating CNNs and Transformers for Multi-Label Classification in 3D CT Imaging

Cervical spine fractures are critical medical emergencies that require urgent diagnosis. The challenge of medical imaging’s complexity and time-intensive nature, underscores the significant need of automation within the medical sector. This study presents a comparative analysis of CNNs and Transformers for Multi-label classification of 3D CT scans. We explore various deep learning models on the RSNA 2022 Cervical Spine Fracture Detection dataset. Notably, both MViT and AutoEncoder demonstrate significant performance enhancements compared to other models in the study.