2026-10-14T21:58:54Z https://www.scidb.cn/oai

10.57760/sciencedb.j00001.01467 2026-01-05T09:33:18Z

2026-01-05 A dataset of gastroenteroscopic polyp object detection doi:10.57760/sciencedb.j00001.01467 en The images in this dataset are sourced from multiple public datasets and data images provided by a certain hospital, including various types and stages of polyps. Collect 5460 images of polyps. The dataset contains three subfolders, each containing the following content: (1) Polyp Images stores original image data in JPG format, with pixel sizes ranging from 300 × 243 to 2559 × 1432. (2) XML Labels stores XML format labels generated by Labelimg software after annotating polyps in images. Each label file name corresponds to the original image and supports custom script conversion formats. (3) TXT Labels stores labels converted from scripts to TXT format files and supports dataset partitioning for YOLO network series. The image resolution ranges from 300 × 243 to 2559 × 1432. The dataset has diverse image acquisition angles and lighting conditions, covering various polyp shapes such as malignant, benign, large and small pixels. The entire dataset has rich samples and strong generalization ability.  polyp; gastroenterological endoscope; target detection Chen Shaobin Chen Jiahui Chen Hongwei Chen Tao Chen Yushu Zhang Jingyuan PUBLIC https://creativecommons.org/licenses/by-nc-sa/4.0/ dataset http://www.doi.org/10.11922/11-6035.nbsdc.2025.0091.zh Science Data Bank