Pedestrian Detection On Caltech

評価指標

Reasonable Miss Rate

評価結果

このベンチマークにおける各モデルのパフォーマンス結果

モデル名
Reasonable Miss Rate
Paper TitleRepository
FRCNN+FPN-Res50+refined feature map+Crowdhuman3.46CrowdHuman: A Benchmark for Detecting Human in a Crowd-
OR-CNN + CityPersons dataset4.1Occlusion-aware R-CNN: Detecting Pedestrians in a Crowd-
ALFNet6.1Learning Efficient Single-stage Pedestrian Detectors by Asymptotic Localization Fitting-
F2DNet (extra data)1.71F2DNet: Fast Focal Detection Network for Pedestrian Detection-
NNNF16.20Pedestrian Detection Inspired by Appearance Constancy and Shape Symmetry-
MS-CNN9.95A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection-
CSP + CityPersons dataset3.8Center and Scale Prediction: Anchor-free Approach for Pedestrian and Face Detection-
F-DNN+SS8.18Fused DNN: A deep neural network fusion approach to fast and robust pedestrian detection-
Zhang et al. *5.1CityPersons: A Diverse Dataset for Pedestrian Detection-
RepLoss + CityPersons dataset4.0Repulsion Loss: Detecting Pedestrians in a Crowd-
RepLoss5.0Repulsion Loss: Detecting Pedestrians in a Crowd-
TA-CNN20.9Pedestrian Detection aided by Deep Learning Semantic Tasks-
F2DNet2.2F2DNet: Fast Focal Detection Network for Pedestrian Detection-
LDCF24.8Local Decorrelation For Improved Pedestrian Detection-
AlexNet23.3Taking a Deeper Look at Pedestrians-
FasterRCNN8.7Is Faster R-CNN Doing Well for Pedestrian Detection?-
Zhang et al.5.8CityPersons: A Diverse Dataset for Pedestrian Detection-
Part-level CNN + saliency and bounding box alignment12.4Part-Level Convolutional Neural Networks for Pedestrian Detection Using Saliency and Boundary Box Alignment-
RPN+BF7.3Is Faster R-CNN Doing Well for Pedestrian Detection?-
NMS-Loss2.92NMS-Loss: Learning with Non-Maximum Suppression for Crowded Pedestrian Detection-
0 of 33 row(s) selected.