Overview
This subset supports research in medical image segmentation, classification, and tumor characterization, providing high-quality annotated ultrasound data to facilitate algorithm development and benchmarking.
This dataset derives from Kaggle and is one of the more widely used public ovarian ultrasound datasets for medical imaging and AI research, particularly for tumor segmentation and classification tasks.
The 2D data represent 1,469 conventional 2D ultrasound images of ovarian tumors from approximately 294 patients.
Each image is paired with pixel-wise semantic annotations that precisely delineate tumor regions, as well as global category labels that classify the tumor type at the image level. This subset supports research in medical image segmentation, classification, and tumor characterization, providing high-quality annotated ultrasound data to facilitate algorithm development and benchmarking.