QueST: Synchronized RGB-D Simulation Sequences for Long-Horizon Tracking

Synchronized RGB-D simulation sequences rendered in SAPIEN from PartNet-Mobility articulated objects. The dataset was created by AnandMayank and accompanies a paper presented at the CAO Workshop at ICLR 2026. It is designed to stress-test long-horizon point tracking algorithms.

Use Cases

• Stress-testing long-horizon point tracking algorithms based on synchronized RGB-D sequences.
• Training models for articulated object pose estimation based on simulation data.
• Benchmarking robotic manipulation algorithms based on simulated PartNet-Mobility objects.
• Developing drift suppression methods for visual tracking based on persistent query semantics.

Strengths

• Dataset is synchronized, providing aligned RGB and depth data streams.
• Data is rendered in the SAPIEN simulation environment, offering controlled conditions.
• Focuses on articulated objects from the PartNet-Mobility benchmark.

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 file formats are unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Simulation sequences rendered in SAPIEN from PartNet-Mobility objects.

Freshness

Last updated 2026-06-06 23:40:18; freshness should be verified.