Researchers have developed TokenSTFormer, a novel model designed for screening Adolescent Idiopathic Scoliosis (AIS) by analyzing gait videos. This model utilizes tokenized spatial and temporal attention mechanisms to improve feature representation and convergence. The ScoliGait dataset, containing 1,516 gait clips and X-ray records, was used to train and evaluate TokenSTFormer, which achieved an accuracy of 0.79, outperforming standard Vision Transformer models. AI
IMPACT This research demonstrates the potential of AI in medical screening, offering a more scalable and objective method for detecting scoliosis.
RANK_REASON The cluster describes a new AI model presented in an academic paper with performance metrics. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
