Skyfall-GS: Evaluation Videos for 3D Urban Scene Synthesis

Skyfall-GS Eval contains evaluation videos for the Skyfall-GS framework, which synthesizes immersive 3D urban scenes from satellite imagery. The dataset was created by author jayinnn and was last updated on March 18, 2026. It is hosted on the Hugging Face platform.

Use Cases

• Evaluate the visual quality of synthesized 3D urban scenes based on the video demonstrations.
• Compare the output of the Skyfall-GS framework against other 3D reconstruction methods based on the provided videos.
• Assess the realism and immersion of city-block scale 3D scenes generated from satellite data.
• Use the videos for qualitative analysis in research papers or project presentations.

Strengths

• Dataset is directly linked to a published research paper (Skyfall-GS) with a project page and arXiv entry.
• Last update timestamp is precisely recorded as 2026-03-18 10:11:24.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.

Provenance

Source

jayinnn on Hugging Face, associated with the Skyfall-GS research project.

Collection Method

Likely contains videos generated as part of the evaluation process for the Skyfall-GS framework.

Freshness

Last updated 2026-03-18 10:11:24; freshness should be verified.