Related Experiment Video
Updated: Jul 30, 2025
Quantitative Fundus Autofluorescence for the Evaluation of Retinal Diseases
Published on: March 11, 2016
MSHF: A Multi-Source Heterogeneous Fundus (MSHF) Dataset for Image Quality Assessment.
Kai Jin1, Zhiyuan Gao1, Xiaoyu Jiang2
1Eye Center, The Second Affiliated Hospital, School of Medicine, Zhejiang University, Zhejiang Provincial Key Laboratory of Ophthalmology, Zhejiang Provincial Clinical Research Center for Eye Diseases, Zhejiang Provincial Engineering Institute on Eye Diseases, Zhejiang, Hangzhou, 310009, China.
This study introduces a large, diverse fundus image quality assessment (IQA) dataset, crucial for improving computer-aided diagnosis in ophthalmology. The MSHF dataset aids in developing standardized medical imaging databases.
Area of Science:
- Ophthalmology
- Medical Imaging
- Computer-Aided Diagnosis
Background:
- Image quality assessment (IQA) is vital for computer-aided diagnosis (CADx) systems.
- Existing fundus IQA datasets are often limited by single-center collection, neglecting device, condition, and environmental variations.
- Standardized, diverse datasets are needed for robust ophthalmic disease diagnosis.
Purpose of the Study:
- To introduce the Multi-Source Heterogeneous Fundus (MSHF) dataset for image quality assessment.
- To address the limitations of existing single-center fundus IQA datasets.
- To facilitate the development of standardized medical image databases for ophthalmic applications.
Main Methods:
- Collected 1302 high-resolution fundus images from diverse sources: color fundus photography (CFP), portable cameras, and ultrawide-field (UWF) imaging.
- Included both normal and pathological images, specifically from diabetic retinopathy patients.
- Image quality was evaluated by three ophthalmologists based on illumination, clarity, contrast, and overall quality.
Main Results:
- The MSHF dataset is one of the largest fundus IQA datasets available.
- Dataset diversity was confirmed through spatial scatter plot visualization.
- Expert ophthalmologists assessed image quality across multiple parameters.
Conclusions:
- The MSHF dataset provides a valuable resource for advancing fundus image quality assessment.
- This work contributes to the creation of standardized medical image databases.
- The dataset will support the development of more reliable ophthalmic CADx systems.
