SR-Ground: Image Quality Grounding for Super-Resolved Content

A dataset accompanying the paper 'SR-Ground: Image Quality Grounding for Super-Resolved Content'. It contains images organized in folders named according to super-resolution method and preset. The dataset was uploaded by author Divotion and last updated on May 21, 2026.

Use Cases

• Training models to assess the perceptual quality of super-resolved images based on the provided image pairs.
• Benchmarking super-resolution algorithms using the grounded quality annotations mentioned in the paper.
• Studying the relationship between specific super-resolution methods and visual artifacts based on the dataset's structure.

Strengths

• Accompanies a published research paper, providing context and methodology.
• Images are organized in a structured folder system based on super-resolution method and preset.

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 is unknown, which may limit suitability assessment.

Provenance

Source

Divotion (author on Hugging Face)

Collection Method

Created as supplementary material for the SR-Ground research paper.

Freshness

Last updated 2026-05-21 02:36:03; freshness should be verified.