Worldrenderer Dataset Test: 27 Synthetic 3D Scenes for Neural Rendering

27 textured 3D scenes rendered for research in novel view synthesis and neural rendering. Each scene includes 401 frames captured along a predefined monocular camera trajectory. The dataset, created by Tengpaz, is a synthetic multi-scene rendering test for 3D-conditioned generative modeling.

Use Cases

• Train novel view synthesis models based on the 401-frame monocular camera trajectories per scene.
• Benchmark neural rendering techniques using the aligned multi-modal data mentioned in the description.
• Develop geometry-aware generation models based on the textured 3D scene data.
• Learn multi-view representations from the synthetic multi-scene structure.
• Build world models for 3D-conditioned generative modeling using the provided scenes.

Strengths

• Contains 27 distinct textured 3D scenes.
• Each scene provides 401 frames rendered from a predefined camera trajectory.
• Designed for multiple advanced research areas including novel view synthesis and neural rendering.

Limitations

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

Provenance

Source

huggingface

Collection Method

Synthetic rendering.

Freshness

Last updated 2026-05-19 09:26:55; freshness should be verified.