2026-10-07T01:39:40Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34033485 2026-09-30T15:42:59Z category_24781 category_24541 category_26275 item_type_3 month_year_09_2026

GastricUS-Seg: an expert-annotated oral contrast-enhanced ultrasound image dataset for gastric cancer lesion segmentation Qingxue Zhao (25144683) Tie Zhou (25144697) Mingyu Mao (24617262) Chenxi Cai (3747667) Jiakang Gao (25144709) Di Wu (23906) Jun Tian (308922) Cancer diagnosis Radiology and organ imaging Biomedical imaging gastric cancer ultrasonography oral contrast-enhanced ultrasound lesion segmentation medical image segmentation deep learning benchmark dataset <p dir="ltr">GastricUS-Seg contains 616 transabdominal B-mode ultrasound images from 104 patients with pathologically confirmed primary gastric cancer, acquired after oral contrast filling of the stomach at a single tertiary hospital in China and organised into eight anonymised acquisition groups (A–H). Every image is paired with a binary pixel-level lesion mask (PNG, 255 = lesion, 0 = background). The release also contains official patient-level five-fold cross-validation splits, de-identified patient and acquisition-group metadata, a log of the burned-in text that was masked during de-identification, benchmark results of thirteen segmentation networks, and the code for loading, integrity checking, evaluation and split generation. File structure and field definitions are given in README.md. The dataset is described in a Data Descriptor submitted to Scientific Data.</p> 2026-09-30T15:42:59Z Dataset Dataset 10.6084/m9.figshare.34033485.v1 https://figshare.com/articles/dataset/GastricUS-Seg_an_expert-annotated_oral_contrast-enhanced_ultrasound_image_dataset_for_gastric_cancer_lesion_segmentation/34033485 CC BY 4.0