A Comprehensive Benchmark Dataset for 2D Organoid Image Segmentation and Object Detection

This dataset is constructed for benchmarking organoid segmentation and object detection in 2D images. It contains carefully annotated images comprising several sub-datasets to ensure diversity, including OrganoidBasic, Lung_organoid, cocoTellu, stomach_intance, and Human_duodenal_biopsy_organoid.The data is intended to support the training, reproduction, and evaluation of computer vision models  in the biomedical domain. It also provides the foundational data for evaluating novel metrics.

本数据集专为二维图像中的类器官（organoid）分割与目标检测基准测试构建。其收录多个经精细标注的图像子数据集以保障数据多样性，涵盖OrganoidBasic、Lung_organoid、cocoTellu、stomach_intance及Human_duodenal_biopsy_organoid。本数据集旨在支撑生物医学领域计算机视觉模型的训练、复现与评估，同时可为新型评估指标的研发提供基础数据支撑。