VGGFace2-HQ: High-Resolution Faces for Academic Editing

VGGFace2-HQ is a high-resolution version of the VGGFace2 dataset created for academic face editing. The dataset was generated using GFPGAN for image restoration and insightface for data preprocessing, including crop and align operations. It is hosted on Hugging Face by RichardErkhov and was last updated on September 25, 2024.

Use Cases

• Training face swapping models based on the high-resolution aligned face images.
• Benchmarking face restoration algorithms using the GFPGAN-enhanced images.
• Conducting academic research on face editing and manipulation with a high-quality dataset.

Strengths

• Dataset is explicitly described as high-resolution, which is a key feature for image editing tasks.
• The generation process is detailed, involving GFPGAN for restoration and insightface for alignment, suggesting a structured pipeline.
• The dataset is hosted on a major platform (Hugging Face) and was updated on 2024-09-25, indicating recent activity.

Limitations

• Description metadata is limited; actual data quality, size, and specific contents require manual inspection after download.
• Column-level documentation is absent; field semantics and data structure must be inferred after download.
• Row count, file formats, and license information are unknown, which may limit suitability assessment.

Provenance

Source

Clone of a GitHub repository (https://github.com/NNNNAI/VGGFace2-HQ), with the original VGGFace2 dataset as a base.

Collection Method

Images were processed using GFPGAN for restoration and insightface for cropping and alignment.

Freshness

Last updated 2024-09-25 14:32:22; freshness should be verified.