欢迎您访问 最编程 本站为您分享编程语言代码,编程技术文章!
您现在的位置是: 首页

理解目标检测中IOU不同范围(如0.5到0.95)的含义与应用

最编程 2024-08-01 20:44:10
...

目标检测结果IOU不同取值的意义(0.5:0.95等)

  • IOU=[0.5:0.95], IOU = 0.5, IOU=0.75等等

IOU=[0.5:0.95], IOU = 0.5, IOU=0.75等等

1.Unless otherwise specified, AP and AR are averaged over multiple Intersection over Union (IoU) values. Specifically we use 10 IoU thresholds of .50:.05:.95. This is a break from tradition, where AP is computed at a single IoU of .50 (which corresponds to our metric APIoU=.50). Averaging over IoUs rewards detectors with better localization.

2.AP is averaged over all categories. Traditionally, this is called “mean average precision” (mAP). We make no distinction between AP and mAP (and likewise AR and mAR) and assume the difference is clear from context.

也就是说:它是直接把mAP当成AP,然后再把IOU值大于0.5的AP(mAP),以0.05的增量,到0.95,也就是以(0.5, 0.55, 0.6, 0.65, 0.7, 0.75, 0.8, 0.85, 0.9, 0.95)IOU值的AP(mAP)的平均值当成AP(at IoU=.50:.05:.95),mscoco进行这样IOU增量平均的考虑可能是:只以0.5IOU为阀值的时候不一定就是更好的模型,可能仅仅在0.5阀值表现的很好,在0.6,0.7…阀值表现的很差,为了更好地评估整体模型的准确度,从而计算一个模型在各个IOU值的AP(mAP),取平均值。

目前目标检测的输出结果展示均为0.5:0.95的IOU递增结果,以此来评判网络好坏更加准确,平衡

推荐阅读