Dataset Overview
Definition:Endoscope video and key frame reportsis provided byChangsha Langhui Information Technology Co., Ltd.Medical data sets provided. Covers 5 million endoscope videos and key frameshots and corresponding diagnostic reports, which correspond to frame labels and can be used for endoscope AI full-process training.
| Dataset Name | Endoscope video and key frame report data sets |
| Total Data Volume | Five million cases of endoscope video. |
| Data Type | Endoscope video + Key frame image + text diagnostic report |
| Spear position, size (mm), morphology (Paris/Manda) | |
| Covered Body Parts | Esophagus, stomach, 12-finger intestine, colon, rectum. |
| License Type | Sublicensing |
Core Data Fields
Each piece of data contains complete structured fields, covering all requirements from basic metadata to professional analysis.
| Field Category | Field Count | Typical Fields |
|---|---|---|
| 6 | Video files (MP4/AVI), frame, resolution, check length, endoscopy type | |
| Key Frame Image | 8 | Key frameshots, corresponding time stamp, frame description, parametric frame for diseased areas |
| Diagnostic reports | 10 | Examinations, position/size/quantity/speculation, results of active examinations, if any, pathological association |
| Patient Information | 4 | Age, gender, date of examination, consultation room (de-identification) |
Data Sample Showcase
The following is a de-identified structured JSON sample illustrating the typical data format.
{
"study_id": "ENDO-2023-0004567",
"meta": {"age_group": "50-60","gender": "male","scope_type": "sample value"},
"findings": [{"polyp_id": 1,"location": "sample value","size_mm": 8,"paris_type": "Ip","yamada": 4}],
"keyframes": [{"timestamp": "03:22","desc": "sample(IpCN)","bbox": [320,240,480,400]}],
"diagnosis": "sample(IpCN),sample","pathology_ref": "PATH-2023-004567"
}AI Application Scenarios
Indigestion Channel AI-assisted screening
An endoscope video training depth learning model to enable real-time detection and positioning of salivated meat/early cancer.
Swallow-Specific Smart Recognition
Key frame labels are used to train the tectonic model (Paris/Madas)
Insight control automation
Video data training internal mirror operation quality control AI, automatic assessment of inspection integrity and operating specifications.
Blind zone monitoring and alerts
Training in blind monitoring models to alert endoscopy physicians to leak-prone areas.
Auto Generate Endoser Report
Video-Key Frame-Report Linked Data Training End-to-end Imagination Report Generation Model.
Endoscope AI teaching system
A rich video data base was used for the training of the ISI teaching system.
Frequently asked questions
Need a customized data solution?
Changsha Langhui Information Technology Co., Ltd. The DataAssetsAPI platform is dedicated to providing high-quality, compliant data assets for AI companies and research institutions. It supports flexible combination by dimension and by quantity scale.
Data Sample Preview · Customized Data Solutions · Professional Technical Support