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

Figure 4

Structure of architecture 2: M-Net encoder+ASPP decoder.