Home/Datasets/CatScreen: A Large MultiModal Benchmark Dataset for Cataract Screening

CatScreen: A Large MultiModal Benchmark Dataset for Cataract Screening

CatScreen is a large multimodal benchmark dataset of 18,640 slit-lamp images from 2,251 subjects, developed by IIT Jodhpur for AI-based cataract screening research, supporting tasks such as diagnosis, severity grading, subtype identification, and image quality assessment across clean, noisy, and unlabelled subsets.

About Dataset

CatScreen is a large multimodal benchmark dataset for AI-based cataract screening, developed by IIT Jodhpur. It contains 18,640 slit-lamp images from 2,251 subjects, captured using a portable Remidio PSL-D20 device. The dataset is split into clean (labelled), noisy, and unlabelled subsets, supporting tasks including image quality assessment, diagnosis, cataract subtype identification, and severity grading. It also includes subject-level metadata and anatomical region annotations. Available for research and educational use via a license agreement.

Purpose of Dataset

To Support The Development Of Robust, Clinically Relevant Ai Systems For Slit-lamp Based Cataract Screening.

Dataset Metadata

License

Other

Geographical coverage

India

Sector

Healthcare, Wellness and Family Welfare

Author

Image Analysis and Biometrics Lab, IIT Jodhpur

Source organisation

ICMR-National Institute for Research in Digital Health and Data Science

Uploaded by

Richa Singh

Data Quality Score (Beta)

Dataset type

Unstructured

Frequency

Static

Time Granularity

Static

Year range

N.A.

Date & Time

11/06/26 09:54:34

Visibility

Open

Hosted / Redirected

Redirected

Data Type

Primary

If Redirection which source

Image Analysis And Biometrics Lab, Iit Jodhpur

Data Collection Method

Clinical Slit-lamp Imaging Using A Remidio Portable Slit Lamp Psl-d20 Device, Acquired Across Real-world Screening Conditions With Variability In Image Quality, Illumination, And Acquisition Settings.