AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

TerraDiT-Ω: Unified Spatial Control for Satellite Image Synthesis with Any Geospatial Primitive

Brian Wei, Srikumar Sastry, Daniel Cher, Eric Xing, Nathan Jacobs

ECCV 2026 · Image generation › Autoregressive and latent generation · approach: Compositional scene generation

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add rotation θ to the RBF prior (ablation)GALA (B, 400k) · Ablation splitFID: 23.52 -> 22.41Improved
add spatial geometry field modulation (ablation)GALA (B, 400k) · Ablation splitFID: 22.41 -> 21.97Improved
add RBF Gaussian prior to cross attention (ablation)Unified Primitive Encoder (B, 400k) · Ablation splitFID: 24.08 -> 23.52Improved
add primitives Ω as conditioning (ablation)TerraDiT-Ω (B, 400k) · Ablation splitFID: 27.19 -> 21.97Improved
add global text conditioning (ablation)TerraDiT-Ω (B, 400k) · Ablation splitFID: 24.90 -> 21.97Improved
add geolocation conditioning (ablation)TerraDiT-Ω (B, 400k) · Ablation splitFID: 23.53 -> 21.97Improved
use GALA instead of Gated Self-Attention (ablation)Unified Primitive Encoder (B, 400k) · Ablation splitFID: 22.27 -> 21.97Improved
use GALA instead of Instance Masked Attention (ablation)Unified Primitive Encoder (B, 400k) · Ablation splitFID: 22.31 -> 21.97Improved
richer primitives: points to box to combined Ω (ablation)TerraDiT-Ω-XL · Git-Rand-15kFID: 9.91 -> 9.72 -> 9.25Improved
richer primitives from points to Ω (ablation)TerraDiT-Ω-XL · Git-Rand-15k CASTop-1 accuracy: 76.66 -> 79.36Improved
add x2 synthetic dataTerraDiT-Ω-XL on DIOR · DIOR + RT-DETRmAP@50-95: 56.14 -> 55.90Mixed
add AeroGen synthetic data without filteringRT-DETR on DIOR · DIOR + RT-DETRmAP@50-95: 55.15 -> 47.20Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy