RetailGoods: Real-World Retail Checkout Product Detection Dataset

RetailGoods is a partially released dataset for product-level visual recognition in checkout scenarios. It includes desensitized sample images, product detection annotations, and train/validation/test splits, released under CC-BY-4.0 by jacky wang. The dataset is intended for academic research and verification, with privacy-sensitive information removed.

Use Cases

• Train object detection models based on annotated product images.
• Benchmark product recognition algorithms in real-world retail checkout environments.
• Research privacy-preserving computer vision techniques based on desensitized visual data.

Strengths

• Includes product detection annotations and dataset splits, facilitating model training and evaluation.
• Visual samples have been processed to reduce privacy risks, removing sensitive information like customer identities and timestamps.

Limitations

• The dataset is a partially released subset; the complete raw dataset is not fully available.
• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

jacky wang via figshare

Collection Method

Constructed from real-world retail checkout scenarios, with data desensitized.

Freshness

Last updated 2026-06-04 03:21:53; freshness should be verified.