EndoWLI-NBI: Endoscopic Images with WLI and NBI Modalities

11,971 high-quality endoscopic image frames were collected from 868 patients at Peking Union Medical College Hospital. The dataset includes both white light imaging (WLI) and narrow band imaging (NBI) under magnifying and non-magnifying conditions, categorized into four classes. It was authored by Faguang Wang and last updated in April 2026.

Use Cases

• Classifying endoscopic findings based on the four provided categories.
• Training models for lesion detection using the high-quality WLI and NBI frames.
• Developing image enhancement algorithms by comparing magnifying and non-magnifying conditions.
• Benchmarking multi-modal image analysis techniques on paired WLI and NBI data.

Strengths

• Contains 11,971 high-quality image frames, providing a substantial volume for model training.
• Sourced from a single institution (Peking Union Medical College Hospital), suggesting consistent collection protocols.
• Includes paired imaging modalities (WLI and NBI) and conditions (magnifying/non-magnifying), enabling comparative analysis.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect geographic or institutional bias inherent to its single-hospital source.
• Row count (i.e., number of distinct image files or patient records) is unknown, which may limit suitability assessment.

Provenance

Source

Peking Union Medical College Hospital

Collection Method

Collected endoscopic imaging data.

Freshness

Last updated 2026-04-21 02:29:06; freshness should be verified.

Geography

China (inferred from hospital location)