Sign in to view source links and access this dataset
Description
20,005 person sequences in the GTA-Human dataset and 35,352 sequences in GTA-Human II, released in 2022 and 2024 respectively, provide synthetic data for 3D human recovery. The datasets were created by caizhongang and are hosted on HuggingFace. They are associated with research published in TPAMI 2024.
Use Cases
Training 3D human mesh recovery models based on synthetic color images.
Benchmarking human pose estimation algorithms on large-scale synthetic sequences.
Developing models for human motion analysis using the sequence data.
Strengths
GTA-Human II contains 35,352 person sequences, providing a large-scale resource.
The datasets are associated with peer-reviewed research published in TPAMI 2024.
Updates show active maintenance, with a recent release on HuggingFace in October 2024.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count for individual sequences is unknown, which may limit suitability assessment.
Data may reflect bias inherent to the synthetic generation process from a video game engine.
Provenance
Source
huggingface
Collection Method
Synthetically generated, likely from a video game engine (GTA).
Freshness
Last updated 2024-10-04 10:13:36
License is unknown; terms of use must be verified before application.