Sign in to view source links and access this dataset
Description
SEAGULL-100w is a large-scale synthetic dataset for no-reference image quality assessment focused on regions of interest. It was created by Zevin2023 and includes images with six distortion types: blur, sharpness, exposure, contrast, colorfulness, and compression. The dataset was last updated on the Hugging Face platform in November 2024.
Use Cases
Training no-reference IQA models based on synthetic distortion types like blur and compression.
Benchmarking region-of-interest IQA algorithms based on the described distortion framework.
Fine-tuning vision-language instruction models for image quality evaluation based on the dataset's construction pipeline.
Strengths
Includes six distinct, synthetically generated distortion types as described.
Dataset scale is described as 'large-scale' in the provided description.
Specifically designed for region-of-interest (ROI) based image quality assessment.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count, file formats, and sample data are unknown, which may limit suitability assessment.
Description metadata is limited; actual data quality requires manual inspection after download.
Provenance
Source
Zevin2023 on Hugging Face.
Collection Method
Synthetically generated images with applied distortions, as described in the construction pipeline.
Freshness
Last updated 2024-11-25 07:56:52; freshness should be verified.
License is unknown; users must verify permissions before use.