预测是1 | 预测是0 | |
实际是1 | TP | FN |
实际是0 | FP | TN |
1. precision=TP/(TP+FP)
2. recall=TP/(TP+FN)
3. IoU=(TP+TN)/(TP+TN+FP+FN)
4. mAP(mean average precision)=所有类别的平均精度之和/类别数目
预测是1 | 预测是0 | |
实际是1 | TP | FN |
实际是0 | FP | TN |
1. precision=TP/(TP+FP)
2. recall=TP/(TP+FN)
3. IoU=(TP+TN)/(TP+TN+FP+FN)
4. mAP(mean average precision)=所有类别的平均精度之和/类别数目
转载于:https://www.cnblogs.com/0-618/p/9376310.html