Using Constant Padding, Reflection Padding and Replication Padding with Keras - MachineCurve

ConstantPadding

• 不解释，就是固定一个值Padding就完了。

ReflectionPadding

• 不好解释，看图就完了
• 好处：Possibly, this improves the performance of your model, because padded inputs will still look like the original ones in terms of data distribution (Liu et al., 2018).

ReplicationPadding SymmetricPadding

• 和ReflectionPadding很像，不过Padding的时候选取的镜像是最近的对应行或列