iraqigold/lung-cancer-segmentation-vision

该数据集是一个用于肺癌肿瘤分割的医学放射影像数据集，经过预处理和标准化，可直接用于深度学习模型训练。原始数据来自Kaggle平台的肺癌分割数据集，经过自定义Python流程优化，包括合并分散的文件夹结构、采用唯一可追溯的命名约定，并将所有数据转换为无损PNG格式，确保CT扫描与病理掩码之间的像素级对应。数据集包含/images文件夹（存放肺部2D CT切片）和/masks文件夹（存放对应的肿瘤区域二进制分割掩码），适用于训练U-Net等分割网络，以支持临床自动化诊断，帮助AI模型精确理解肺结节的空间形态、大小和边界。

This dataset is a preprocessed and standardized version of medical radiography images for lung cancer tumor segmentation, ready for direct use in deep learning model training. The raw data was sourced from the Kaggle Lung Cancer Segmentation dataset, refined through a custom Python pipeline that flattened hierarchical folders, implemented a unique traceable naming convention, and converted all data to lossless PNG format for exact pixel mapping between CT scans and pathological masks. It includes an /images folder (containing 2D CT lung slices) and a /masks folder (containing corresponding binary ground-truth tumor masks), optimized for training segmentation networks like U-Net to aid in automated clinical diagnostics by enabling AI models to understand the precise spatial morphology, size, and boundaries of lung nodules.