gushiyi27的博客

top1-Accuracy,top5-Accuracy举例精析符号定义：TP(True Positive):标签为正样本且分类为正样本的数目FN(False Negative):标签为正样本而分类为负样本的数目FP(False Positive):标签为负样本而分类为正样本的数目TN(True Negative):标签为负样本且分类为负样本的数目Accuracy=TP+FNTP+FN+FP+TNAccuracy =\dfrac{TP+FN}{TP+FN+FP+TN}Accur

UNet Family Ideasmodule汇总：DAC(Dense Atrous Convolution module)RMP(residual multi-kernel pooling)AAM(Augmented Attention Module)CAM(channel attention module)SAM(sequential channel attention and spatial atte