ResearchBenchmarksStepwise validity estimation on Sofa ScenarioFollow99.7Mean Inclusion Rate (Final)GMM99.38899.46999.5599.631Mar 4, 2026Evaluation ResultsMethodMethodLinksMean Inclusion Rate (Final)Mean Inclusion Rate (All)Final Step Inclusion Rate >= 95%All Step Inclusion Rate >= 95%Path Compactness (Final)GMMModel=GMMModel=GMM2026.0399.799.310097.810.5EllipModel=EllipModel=Ellip2026.0399.498.896.793.79.48