RECORD
SIGN: Spatial-information Incorporated Generative Network for Generalized Zero-shot Semantic Segmentation
Computer Vision and Pattern Recognition · Submitted 27 Aug 2021
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim We systematically study the proposed Relative Positional Encoding and Annealed Self-Training in a comprehensive experimental evaluation, and our empirical results confirm the effectiveness of our method on three benchmark datasets.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
Unlike conventional zero-shot classification, zero-shot semantic segmentation predicts a class label at the pixel level instead of the image level. When solving zero-shot semantic segmentation problems, the need for pixel-level prediction with surrounding context motivates us to incorporate spatial information using positional encoding. We improve standard positional encoding by introducing the concept of Relative Positional Encoding, which integrates spatial information at the feature level and can handle arbitrary image sizes. Furthermore, while self-training is widely used in zero-shot semantic segmentation to generate pseudo-labels, we propose a new knowledge-distillation-inspired self-training strategy, namely Annealed Self-Training, which can automatically assign different importance to pseudo-labels to improve performance. We systematically study the proposed Relative Positional Encoding and Annealed Self-Training in a comprehensive experimental evaluation, and our empirical results confirm the effectiveness of our method on three benchmark datasets.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | Unlike conventional zero-shot classification, zero-shot semantic segmentation predicts a class label at the pixel level instead of the image level. | background 1.00 | 0.05 | 0.03 | asserted 1.00 |
| s1 | When solving zero-shot semantic segmentation problems, the need for pixel-level prediction with surrounding context motivates us to incorporate spatial information using positional encoding. | background 0.60 | 0.24 | 0.03 | asserted 0.95 |
| s2 | We improve standard positional encoding by introducing the concept of Relative Positional Encoding, which integrates spatial information at the feature level and can handle arbitrary image sizes. | method 0.56 | 0.92 | 0.06 | asserted 0.99 |
| s3 | Furthermore, while self-training is widely used in zero-shot semantic segmentation to generate pseudo-labels, we propose a new knowledge-distillation-inspired self-training strategy, namely Annealed Self-Training, which can automatically assign different importance to pseudo-labels to improve performance. | contribution 0.75 | 0.90 | 0.07 | asserted 0.93 |
| s4 | We systematically study the proposed Relative Positional Encoding and Annealed Self-Training in a comprehensive experimental evaluation, and our empirical results confirm the effectiveness of our method on three benchmark datasets. | result 0.84 | 0.97 | 0.84 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original abstract
Unlike conventional zero-shot classification, zero-shot semantic segmentation predicts a class label at the pixel level instead of the image level. When solving zero-shot semantic segmentation problems, the need for pixel-level prediction with surrounding context motivates us to incorporate spatial information using positional encoding. We improve standard positional encoding by introducing the concept of Relative Positional Encoding, which integrates spatial information at the feature level and can handle arbitrary image sizes. Furthermore, while self-training is widely used in zero-shot semantic segmentation to generate pseudo-labels, we propose a new knowledge-distillation-inspired self-training strategy, namely Annealed Self-Training, which can automatically assign different importance to pseudo-labels to improve performance. We systematically study the proposed Relative Positional Encoding and Annealed Self-Training in a comprehensive experimental evaluation, and our empirical results confirm the effectiveness of our method on three benchmark datasets.
Human review
How the literature uses it
1 arXiv paper cite this in 2 sentences
Papers citing it
Compares with it 1 paper
-
Specifically, ZegFormer outperforms Joint [ref] by 31 points in the mIoU of unseen classes on PASCAL VOC [ref] , and outperforms SIGN [ref] by 18 points in the mIoU of unseen classes on COCO-Stuff [ref] .
Decoupling Zero-Shot Semantic Segmentation · Comparison with the State-of-the-art · 1.00
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.