WRSWorld Research StateContribute
WORLD RESEARCH STATE

DeepLab family: atrous context modeling and boundary refinement

Contribution grade

Semantic image segmentation · Human-designed research method or resource

Reference score180,000Provisional editorial estimate of this exact contribution; its grade is assessed separately.

What changed

Adapted dense convolutional features with atrous sampling and CRF boundary refinement; later added multi-rate spatial context pooling, image-level context and an encoder–decoder with atrous separable convolutions.

Scientific scope

Adapted dense convolutional features with atrous sampling and CRF boundary refinement; later added multi-rate spatial context pooling, image-level context and an encoder–decoder with atrous separable convolutions. Atrous convolution inherits wavelet/signal-processing and earlier neural precedents. Versions do not all use the same CRF: v3 and v3+ report results without that post-processing. Dense sampling, extra context and decoder refinement are version-specific contributions, not four independent inventions of segmentation.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Atrous convolution inherits wavelet/signal-processing and earlier neural precedents. Versions do not all use the same CRF: v3 and v3+ report results without that post-processing. Dense sampling, extra context and decoder refinement are version-specific contributions, not four independent inventions of segmentation.

Contributors

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction