Title | ||
---|---|---|
Stable self-attention adversarial learning for semi-supervised semantic image segmentation |
Abstract | ||
---|---|---|
An overview of the proposed system for semi-supervised semantic image segmentation, where the segmentation network G outputs a class probability map, SA represents the self-attention modules, SN represents the application of the spectral normalization technique, the discriminator network D outputs a confidence map, Lce is the spatial multi-class cross entropy loss based on the ground truth label map, Ladv is the adversarial loss of D, and Lsemi is the masked cross entropy loss. We use the loss LD to train the discriminator based on the full convolutional network. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1016/j.jvcir.2021.103170 | Journal of Visual Communication and Image Representation |
Keywords | DocType | Volume |
41A05,41A10,65D05,65D17 | Journal | 78 |
ISSN | Citations | PageRank |
1047-3203 | 0 | 0.34 |
References | Authors | |
0 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jia Zhang | 1 | 0 | 0.34 |
Zhixin Li | 2 | 12 | 19.62 |
Canlong Zhang | 3 | 5 | 8.55 |
Huifang Ma | 4 | 0 | 1.35 |