SPADE: spatial layout modulation for semantic image synthesis

What changed

Conditioned normalized feature activations with spatially varying scale and bias predicted from semantic layouts, retaining layout information throughout a generator and supporting optional style conditioning.

Scientific scope

Conditioned normalized feature activations with spatially varying scale and bias predicted from semantic layouts, retaining layout information throughout a generator and supporting optional style conditioning. Generalizes earlier conditional normalization and inherits multi-scale adversarial/feature losses. A plausible image need not be a faithful depiction of a real scene. Taesung Park's star denotes internship, not an equal-contribution marker.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Generalizes earlier conditional normalization and inherits multi-scale adversarial/feature losses. A plausible image need not be a faithful depiction of a real scene. Taesung Park's star denotes internship, not an equal-contribution marker.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

• Semantic Image Synthesis with Spatially-Adaptive Normalization

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