A Deep Learning-Based Semantic Segmentation Architecture for Autonomous Driving Applications

Figure 2

Structure of architecture 1: M-Net encoder+PSP decoder.