Morphological Decoupling-Based Skeletal Classification for Clinical Assessment of Malocclusion
TeethGNN is a novel graph-based framework that combines CBCT image features with morphological information to accurately grade malocclusion skeletal classes. The system uses separate pathways for image and morphology, fuses them into a multimodal graph, and employs advanced calibration techniques to achieve high accuracy, serving as a decision-support tool for clinical assessment.