Long, Shelhamer and Darrell: fully convolutional networks for semantic segmentation

What changed

Converted classifiers into fully convolutional pixel predictors and fused coarse and fine features via skip connections.

Scientific scope

The record is the exact three-author CVPR 2015 paper and its reported PASCAL VOC, NYUDv2 and SIFT Flow results. Earlier dense-prediction work, U-Net, later FCN revisions and other segmentation architectures remain separate.

Evidence

The linked sources identify this contribution and its scope. WRS has not independently reconstructed the scientific claims. The record is the exact three-author CVPR 2015 paper and its reported PASCAL VOC, NYUDv2 and SIFT Flow results. Earlier dense-prediction work, U-Net, later FCN revisions and other segmentation architectures remain separate.

Contributors

• Jonathan Long · SCIENCE: Named source author of the scoped work; individual effort shares are not measured.
• Evan Shelhamer · SCIENCE: Named source author of the scoped work; individual effort shares are not measured.
• Trevor Darrell · SCIENCE: Named source author of the scoped work; individual effort shares are not measured.

• CVF record for Fully Convolutional Networks

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