Skip to main content
. 2022 Aug 19;22(16):6226. doi: 10.3390/s22166226

Figure 3.

Figure 3

The process of our method. The red dotted box represents the model at the recognition stage. The model allows for network interaction during training by establishing information communication among a set of parallel DeeplabV3+ networks. E and D represent the encoder network and decoder network of DeeplabV3+, respectively.