Matches in SemOpenAlex for { <https://semopenalex.org/work/W4226367174> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4226367174 abstract "The crux of semi-supervised semantic segmentation is to assign adequate pseudo-labels to the pixels of unlabeled images. A common practice is to select the highly confident predictions as the pseudo ground-truth, but it leads to a problem that most pixels may be left unused due to their unreliability. We argue that every pixel matters to the model training, even its prediction is ambiguous. Intuitively, an unreliable prediction may get confused among the top classes (i.e., those with the highest probabilities), however, it should be confident about the pixel not belonging to the remaining classes. Hence, such a pixel can be convincingly treated as a negative sample to those most unlikely categories. Based on this insight, we develop an effective pipeline to make sufficient use of unlabeled data. Concretely, we separate reliable and unreliable pixels via the entropy of predictions, push each unreliable pixel to a category-wise queue that consists of negative samples, and manage to train the model with all candidate pixels. Considering the training evolution, where the prediction becomes more and more accurate, we adaptively adjust the threshold for the reliable-unreliable partition. Experimental results on various benchmarks and training settings demonstrate the superiority of our approach over the state-of-the-art alternatives." @default.
- W4226367174 created "2022-05-05" @default.
- W4226367174 creator A5000432967 @default.
- W4226367174 creator A5022596740 @default.
- W4226367174 creator A5037249009 @default.
- W4226367174 creator A5041282060 @default.
- W4226367174 creator A5047303618 @default.
- W4226367174 creator A5049852063 @default.
- W4226367174 creator A5052481491 @default.
- W4226367174 creator A5064820015 @default.
- W4226367174 creator A5065586051 @default.
- W4226367174 date "2022-03-08" @default.
- W4226367174 modified "2023-10-14" @default.
- W4226367174 title "Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels" @default.
- W4226367174 doi "https://doi.org/10.48550/arxiv.2203.03884" @default.
- W4226367174 hasPublicationYear "2022" @default.
- W4226367174 type Work @default.
- W4226367174 citedByCount "0" @default.
- W4226367174 crossrefType "posted-content" @default.
- W4226367174 hasAuthorship W4226367174A5000432967 @default.
- W4226367174 hasAuthorship W4226367174A5022596740 @default.
- W4226367174 hasAuthorship W4226367174A5037249009 @default.
- W4226367174 hasAuthorship W4226367174A5041282060 @default.
- W4226367174 hasAuthorship W4226367174A5047303618 @default.
- W4226367174 hasAuthorship W4226367174A5049852063 @default.
- W4226367174 hasAuthorship W4226367174A5052481491 @default.
- W4226367174 hasAuthorship W4226367174A5064820015 @default.
- W4226367174 hasAuthorship W4226367174A5065586051 @default.
- W4226367174 hasBestOaLocation W42263671741 @default.
- W4226367174 hasConcept C106301342 @default.
- W4226367174 hasConcept C114614502 @default.
- W4226367174 hasConcept C119857082 @default.
- W4226367174 hasConcept C121332964 @default.
- W4226367174 hasConcept C124101348 @default.
- W4226367174 hasConcept C146849305 @default.
- W4226367174 hasConcept C153180895 @default.
- W4226367174 hasConcept C154945302 @default.
- W4226367174 hasConcept C160633673 @default.
- W4226367174 hasConcept C199360897 @default.
- W4226367174 hasConcept C33923547 @default.
- W4226367174 hasConcept C41008148 @default.
- W4226367174 hasConcept C42812 @default.
- W4226367174 hasConcept C43521106 @default.
- W4226367174 hasConcept C62520636 @default.
- W4226367174 hasConcept C89600930 @default.
- W4226367174 hasConceptScore W4226367174C106301342 @default.
- W4226367174 hasConceptScore W4226367174C114614502 @default.
- W4226367174 hasConceptScore W4226367174C119857082 @default.
- W4226367174 hasConceptScore W4226367174C121332964 @default.
- W4226367174 hasConceptScore W4226367174C124101348 @default.
- W4226367174 hasConceptScore W4226367174C146849305 @default.
- W4226367174 hasConceptScore W4226367174C153180895 @default.
- W4226367174 hasConceptScore W4226367174C154945302 @default.
- W4226367174 hasConceptScore W4226367174C160633673 @default.
- W4226367174 hasConceptScore W4226367174C199360897 @default.
- W4226367174 hasConceptScore W4226367174C33923547 @default.
- W4226367174 hasConceptScore W4226367174C41008148 @default.
- W4226367174 hasConceptScore W4226367174C42812 @default.
- W4226367174 hasConceptScore W4226367174C43521106 @default.
- W4226367174 hasConceptScore W4226367174C62520636 @default.
- W4226367174 hasConceptScore W4226367174C89600930 @default.
- W4226367174 hasLocation W42263671741 @default.
- W4226367174 hasOpenAccess W4226367174 @default.
- W4226367174 hasPrimaryLocation W42263671741 @default.
- W4226367174 hasRelatedWork W2136485282 @default.
- W4226367174 hasRelatedWork W2157822554 @default.
- W4226367174 hasRelatedWork W2507402573 @default.
- W4226367174 hasRelatedWork W2546871836 @default.
- W4226367174 hasRelatedWork W2565015337 @default.
- W4226367174 hasRelatedWork W2783482906 @default.
- W4226367174 hasRelatedWork W2904499449 @default.
- W4226367174 hasRelatedWork W2907667403 @default.
- W4226367174 hasRelatedWork W3115043162 @default.
- W4226367174 hasRelatedWork W4206076898 @default.
- W4226367174 isParatext "false" @default.
- W4226367174 isRetracted "false" @default.
- W4226367174 workType "article" @default.