DeepLabv3+: Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation

Spatial pyramid pooling module

Encoder-decoder structure

Parallel atrous convolution—— Atrous Spatial Pyramid Pooling—— ASPP

Atrous/Dilated convolution: 增大感受野,但是可以不改变图像输出特征图的尺寸

Depthwise separable convolution: MoblieNet 可减少计算量

Xception: Inception的升级

没啥意思感觉，到处拼拼凑凑