ClothingADC: A Web-Crawled Fashion Image Dataset for Computer Vision

Clothing-ADC is a dataset constructed using Automatic Dataset Construction (ADC) methods, as described in a paper by authors from UC Santa Cruz, HKUST(GZ), and other institutions. The dataset is hosted on Hugging Face and was last updated on March 23, 2026. Platform tags indicate it is a web-crawled collection focused on fashion and clothing images, likely containing millions of samples.

Use Cases

• Training image classification models based on clothing categories mentioned in the description.
• Benchmarking model robustness based on the 'Noisy Labels' and 'Class Imbalance' tags.
• Studying data curation and automatic dataset construction methods as referenced in the paper title.
• Developing feature extraction techniques for fashion and apparel images.

Strengths

• Dataset is associated with a formal research paper on Automatic Dataset Construction (ADC).
• Platform tags suggest a scale in the '1 Mn10 M' (1 million to 10 million) range.
• Last updated on 2026-03-23, indicating recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic or source bias inherent to its web-crawled origin.

Provenance

Source

Constructed by authors from UC Santa Cruz, HKUST(GZ), UC Davis, SUSTech, Zhejiang University, Yale University, Carnegie Mellon University, Nanyang Technological University, and others.

Collection Method

Likely created via web crawling and automated curation, as suggested by the paper title 'Automatic Dataset Construction (ADC): Sample Collection, Data Curation, and Beyond'.

Time Range

null

Freshness

Last updated 2026-03-23 04:36:56.

Geography

Platform tags indicate a 'us' region, but specific coverage is unknown.