Pixel-Level Tear Meniscus Segmentation Dataset with Multimodal Imaging

The dataset introduced is a multimodal, pixel-level tear meniscus segmentation dataset. It contains 1693 colourful modal images and 1739 infrared modal images, collected from five centres. Each image is accompanied by segmentation labels meticulously annotated by professional ophthalmologists. This high-quality, multimodal dataset can effectively support the development of AI models, enabling more accurate segmentation and assessment of tear meniscus height for dry eye diagnosis and treatment research.