混合采样下多级特征聚合的视频目标检测算法

秦思怡,盖绍彦,达飞鹏

Video object detection algorithm based on multi-level feature aggregation under mixed sampler

Siyi QIN,Shaoyan GAI,Feipeng DA

表 2 不同算法在ImageNet VID验证集上的实验结果

Tab.2 Experiment results of different algorithms on ImageNet VID verification set

网络模型 主干网络 t/ms AP50/% FGFA[1] ResNet-101 104.2 76.3 MEGA[10] ResNet-101 230.4 82.9 VOD-MT[30] VGG-16 — 73.2 LSTS[11] ResNet-101 43.5 77.2 TIAM[31] ResNet-101 — 74.9 QueryProp[7] ResNet-50 21.9(T) 80.3 SALISA[32] EfficientNet-B3 — 75.4 YOLOX-S[20] Modified CSP v5 9.4 69.5 MMNet Modified CSP v5 11.5 77.8