Retinal Finding and Diagnosis Consistency Evaluation Dataset

#inconsistency detection #clinical finding validation #diagnosis consistency checking #retinal disease diagnosis #ophthalmic medical record review #clinical finding-diagnosis consistency assessment

5840
0.3GB
JSON
CATL
MOBIUSI INC

Updated:2026-11-04

AI Analysis & Value Prop

This evaluation dataset provides JSON records of structured clinical findings and diagnoses related to the retina, macula, and vitreous, with labels for consistency, conflict type, and explanation. It requires no image files as input and supports evaluation of a model's ability to detect contradictions within clinical records, as well as ophthalmic text quality control and model validation.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
conflict_type	string	Categorizes the conflict between findings and diagnosis; use none when they are consistent.
is_consistent	boolean	Indicates whether the documented diagnosis is consistent with the structured clinical findings.
clinical_record	object	Structured clinical record containing retinal findings and diagnosis information.
conflict_explanation	string	Briefly explains the mismatch between findings and diagnosis, or why they are consistent.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: Each record includes structured clinical findings and diagnosis information, along with a consistency label, conflict type, and brief explanation. The records focus on the retina, macula, and vitreous.

Are retinal images required for evaluation?: No. The benchmark uses structured clinical text in JSON format as input, making it suitable for evaluating whether models can judge logical consistency between documented findings and diagnoses.

Which tasks is the dataset suitable for?: It can support clinical record consistency checking, diagnosis contradiction detection, conflict type classification, and ophthalmic text quality control. It can also serve as an automated evaluation benchmark for medical language models.

How can I access the data, and what are the licensing terms?: Check the dataset product page for available access methods, licensing scope, and usage restrictions. For clinical or research projects, follow the stated license terms and applicable privacy requirements.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Retinal Finding and Diagnosis Consistency Evaluation Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/fb09b6d4630793f362774f60b870f788},
  urldate={2026-10-22},
  keywords={retinal dataset,ophthalmic clinical records,diagnosis consistency benchmark,retinal finding analysis,medical text quality control,diagnosis contradiction detection,macular record data,vitreous clinical data},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Scene

Domain

Modality

Task