【数据增强】CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features

x是图片，y是图片对应的label，通过合成两张训练图片a和b，生成新的训练样本。

M是二值mask，大小与原图片一致。lam来源于Beta(alp, alp)分布。Alp设置为1，lam取自均匀分布（0，1）。设置bbox B，它的坐标是：

裁剪图片a中的区域B，用图片b中的区域B填充。Mask M的长宽比例和原图片的一致。Bbox坐标通过下式得到：

具体算法如下：

.

结果：

On ImageNet classification, applying CutMix to ResNet-50 and ResNet101 brings +2.28% and +1.70% top-1 accuracy improvements. On CIFAR classification, CutMix significantly improves the performance of baseline by +1.98% leads to the state-of-the-art top-1 error 14.47%