Stacked Homography Transformations for Multi-View Pedestrian Detection

Benchmark Model Rank Results
multiview-detection-on-citystreetSHOTMODA (2m): 53.5MODP (2m): 72.4Precision (2m): 91.0
multiview-detection-on-cvcsSHOTMODA (1m): 45.0MODP (1m): 77.4Precision (1m): 83.6
multiview-detection-on-multiviewxSHOTMODA: 88.3MODP: 82.0Recall: 91.5
multiview-detection-on-wildtrackSHOTMODA: 90.2MODP: 76.5Recall: 94.0