Semantic Segmentation under a Complex Background for Machine Vision Detection Based on Modified UPerNet with Component Analysis Modules

Figure 9

Optimized architecture with the component analysis module. (a) Replace with to optimize the module. (b) Analyze to optimize the module. (c) Replace with and analyze to optimize the module.

(a)

(b)

(c)