A Real-Time Image Semantic Segmentation Method Based on Multilabel Classification

Figure 14

Results of the improved ESPNet model on self-built dataset: (a) original image; (b) segmented image; (c) fused image.

(a)

(b)

(c)