ColonCancerHSI
收藏资源简介:
ColonCancerHSI数据集由鲁德·博斯科维奇研究所创建,包含27张高光谱图像,这些图像来自14名患有结肠腺癌并转移到肝脏的患者。数据集的图像在450至800 nm的波长范围内采集,分辨率为1 nm,图像大小为1384×1035像素。数据集的创建过程包括由三位病理学家进行的像素级标注,旨在验证术中肿瘤切除的像素级分类。该数据集主要应用于计算病理学领域,旨在解决术中肿瘤边界难以识别的问题,提高肿瘤切除的准确性。
The ColonCancerHSI dataset was developed by the Ruđer Bošković Institute. It contains 27 hyperspectral images sourced from 14 patients diagnosed with colon adenocarcinoma and liver metastasis. The images were acquired within the wavelength range of 450 to 800 nm, with a spectral resolution of 1 nm, and each image has a size of 1384 × 1035 pixels. The dataset construction included pixel-level annotations performed by three pathologists, aiming to validate pixel-level classification for intraoperative tumor resection. This dataset is primarily applied in the field of computational pathology, targeting the challenge of difficult intraoperative tumor boundary identification and improving the accuracy of tumor resection.




