Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385804923> ?p ?o ?g. }
Showing items 1 to 90 of
90
with 100 items per page.
- W4385804923 abstract "The field-of-view is an important metric when designing a model for semantic segmentation. To obtain a large field-of-view, previous approaches generally choose to rapidly downsample the resolution, usually with average poolings or stride 2 convolutions. We take a different approach by using dilated convolutions with large dilation rates throughout the backbone, allowing the backbone to easily tune its field-of-view by adjusting its dilation rates, and show that it’s competitive with existing approaches. To effectively use the dilated convolution, we show a simple upper bound on the dilation rate in order to not leave gaps in between the covolutional weights, and design an SEResNeXt inspired block structure that uses two parallel 3 × 3 convolutions with different dilation rates to preserve the local details. Manually tuning the dilation rates for every block can be difficult, so we also introduce a differentiable neural architecture search method that uses gradient descent to optimize the dilation rates. In addition, we propose a lightweight decoder that restores local information better than common alternatives. To demonstrate the effectiveness of our approach, our model RegSeg achieves competitive results on real-time Cityscapes and CamVid datasets. Using a T4 GPU with mixed precision, RegSeg achieves 78.3 mIOU on Cityscapes test set at 37 FPS, and 80.9 mIOU on CamVid test set at 112 FPS, both without ImageNet pretraining." @default.
- W4385804923 created "2023-08-15" @default.
- W4385804923 creator A5076162525 @default.
- W4385804923 date "2023-06-01" @default.
- W4385804923 modified "2023-10-14" @default.
- W4385804923 title "Rethinking Dilated Convolution for Real-time Semantic Segmentation" @default.
- W4385804923 cites W1913356549 @default.
- W4385804923 cites W2108598243 @default.
- W4385804923 cites W2194775991 @default.
- W4385804923 cites W2549139847 @default.
- W4385804923 cites W2560023338 @default.
- W4385804923 cites W2601564443 @default.
- W4385804923 cites W2752782242 @default.
- W4385804923 cites W2922509574 @default.
- W4385804923 cites W2961666066 @default.
- W4385804923 cites W2963176324 @default.
- W4385804923 cites W2963419596 @default.
- W4385804923 cites W2963855133 @default.
- W4385804923 cites W2963890956 @default.
- W4385804923 cites W2964254867 @default.
- W4385804923 cites W2982083293 @default.
- W4385804923 cites W2987175876 @default.
- W4385804923 cites W3016719260 @default.
- W4385804923 cites W3034355852 @default.
- W4385804923 cites W3034429256 @default.
- W4385804923 cites W3035441881 @default.
- W4385804923 cites W3035531015 @default.
- W4385804923 cites W3035682985 @default.
- W4385804923 cites W3110440461 @default.
- W4385804923 cites W3169865585 @default.
- W4385804923 cites W3172265425 @default.
- W4385804923 cites W3174068320 @default.
- W4385804923 cites W3175630421 @default.
- W4385804923 cites W3196904463 @default.
- W4385804923 doi "https://doi.org/10.1109/cvprw59228.2023.00493" @default.
- W4385804923 hasPublicationYear "2023" @default.
- W4385804923 type Work @default.
- W4385804923 citedByCount "0" @default.
- W4385804923 crossrefType "proceedings-article" @default.
- W4385804923 hasAuthorship W4385804923A5076162525 @default.
- W4385804923 hasConcept C11413529 @default.
- W4385804923 hasConcept C134306372 @default.
- W4385804923 hasConcept C154945302 @default.
- W4385804923 hasConcept C162324750 @default.
- W4385804923 hasConcept C169903167 @default.
- W4385804923 hasConcept C176217482 @default.
- W4385804923 hasConcept C202615002 @default.
- W4385804923 hasConcept C21547014 @default.
- W4385804923 hasConcept C2524010 @default.
- W4385804923 hasConcept C2777210771 @default.
- W4385804923 hasConcept C2780757906 @default.
- W4385804923 hasConcept C3261483 @default.
- W4385804923 hasConcept C33923547 @default.
- W4385804923 hasConcept C41008148 @default.
- W4385804923 hasConcept C45347329 @default.
- W4385804923 hasConcept C50644808 @default.
- W4385804923 hasConcept C89600930 @default.
- W4385804923 hasConceptScore W4385804923C11413529 @default.
- W4385804923 hasConceptScore W4385804923C134306372 @default.
- W4385804923 hasConceptScore W4385804923C154945302 @default.
- W4385804923 hasConceptScore W4385804923C162324750 @default.
- W4385804923 hasConceptScore W4385804923C169903167 @default.
- W4385804923 hasConceptScore W4385804923C176217482 @default.
- W4385804923 hasConceptScore W4385804923C202615002 @default.
- W4385804923 hasConceptScore W4385804923C21547014 @default.
- W4385804923 hasConceptScore W4385804923C2524010 @default.
- W4385804923 hasConceptScore W4385804923C2777210771 @default.
- W4385804923 hasConceptScore W4385804923C2780757906 @default.
- W4385804923 hasConceptScore W4385804923C3261483 @default.
- W4385804923 hasConceptScore W4385804923C33923547 @default.
- W4385804923 hasConceptScore W4385804923C41008148 @default.
- W4385804923 hasConceptScore W4385804923C45347329 @default.
- W4385804923 hasConceptScore W4385804923C50644808 @default.
- W4385804923 hasConceptScore W4385804923C89600930 @default.
- W4385804923 hasLocation W43858049231 @default.
- W4385804923 hasOpenAccess W4385804923 @default.
- W4385804923 hasPrimaryLocation W43858049231 @default.
- W4385804923 hasRelatedWork W2013435028 @default.
- W4385804923 hasRelatedWork W2358941527 @default.
- W4385804923 hasRelatedWork W2890372105 @default.
- W4385804923 hasRelatedWork W2891538076 @default.
- W4385804923 hasRelatedWork W2951064045 @default.
- W4385804923 hasRelatedWork W2969394474 @default.
- W4385804923 hasRelatedWork W2999972254 @default.
- W4385804923 hasRelatedWork W3021946687 @default.
- W4385804923 hasRelatedWork W4220718606 @default.
- W4385804923 hasRelatedWork W4283016402 @default.
- W4385804923 isParatext "false" @default.
- W4385804923 isRetracted "false" @default.
- W4385804923 workType "article" @default.