FIGURE 4 - Federated Onboard-Ground Station Computing With Weakly Supervised Cascading Pyramid Attention Network for Satellite Image Analysis

FIGURE 4: The background bias problem of conventional visual explanation based on top convolution layer. The trained DL model infers the ground-truth label by focusing background context around the target objects.

The background bias problem of conventional visual explanation based on top convolution layer. The trained DL model infers the ground-truth label by focusing background context around the target objects.

Go to figure page

Reference

Federated Onboard-Ground Station Computing With Weakly Supervised Cascading Pyramid Attention Network for Satellite Image Analysis - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/The-background-bias-problem-of-conventional-visual-explanation-based-on-top-convolution_fig2_365127865 [accessed 15 Mar 2025]

Caption

FIGURE 4: The background bias problem of conventional visual explanation based on top convolution layer. The trained DL model infers the ground-truth label by focusing background context around the target objects.

Embed code

<a href="https://www.researchgate.net/figure/The-background-bias-problem-of-conventional-visual-explanation-based-on-top-convolution_fig2_365127865"><img src="https://www.researchgate.net/publication/365127865/figure/fig2/AS:11431281094833894@1667608008911/The-background-bias-problem-of-conventional-visual-explanation-based-on-top-convolution.ppm" alt="The background bias problem of conventional visual explanation based on top convolution layer. The trained DL model infers the ground-truth label by focusing background context around the target objects."/></a>