ResearchBenchmarksLandmark Regression on AFLW unalignedLandmark Regression on AFLW unalignedFollow13.13Mean ErrorOurs11.41223.008534.60546.2015May 26, 2021Evaluation resultsMethodMethodLinksMean ErrorError Rate (%)OursK=8K=82021.0513.13—IMMK=8K=82021.0513.31—Lorenz et al.K=8K=82021.0513.6—SCOPSK=8, saliency prior=trueK=8, saliency prior=true2021.0516.05—SCOPSK=8, saliency prior=falseK=8, saliency prior=false2021.0556.08—DFFNumber of parts (K)=8Number of parts (K)=82019.05—20.42SCOPSNumber of parts (K)=8Number of parts (K)=82019.05—16.54ULDNumber of parts (K)=8Number of parts (K)=82019.05—25.03