Automated objective and marker-free facial grading using photographs of patients with facial palsy.

摘要: PURPOSE An automated, objective, fast and simple classification system for the grading of facial palsy (FP) is lacking. METHODS observational single center study was performed. 4572 photographs 233 patients with unilateral peripheral FP were subjectively rated automatically analyzed applying a machine learning approach including Supervised Descent Method. This allowed an automated all according to House-Brackmann scale (HB), Sunnybrook (SB), Stennert index (SI). RESULTS Median time first assessment 6 days after onset. At examination, median objective HB, total SB, SI grade 3, 45, 5, respectively. The best correlation between subjective seen SB movement score (r = 0.746; r = 0.732, respectively). No agreement found HB [Test symmetry 80.61, df = 15, p < 0.001, weighted kappa = - 0.0105; 95% confidence interval (CI) = - 0.0542 0.0331; p = 0.6541]. Also no (test 166.37, df = 55, p < 0.001) although there nonzero kappa = 0.2670; CI 0.2154-0.3186; p < 0.0001). Based on multinomial logistic regression probability higher scores compared (OR 1.608; 1.202-2.150; p = 0.0014). (ICC = 0.34645). CONCLUSIONS Automated delivered fair global regional data motor function use in clinical routine trials.