医学影像数据集|息肉检测数据集

<p>Polyp-Gen数据集是一个专注于内窥镜息肉图像生成的真实、多样化数据集，旨在提升息肉检测模型的泛化能力。数据集共包含55,883个高质量内窥镜图像帧，其中29,640个为息肉帧，26,243个为非息肉帧，所有图像均经过严格筛选，剔除模糊、反射和重影等低质量样本。其特点是图像质量高、类别分布均衡、场景多样，适用于医学图像分析中的息肉检测、分割及数据增强任务，可广泛应用于结肠镜检查辅助诊断系统的开发与评估。</p>Polyp-Gen is a real-world and diverse dataset focused on endoscopic polyp image generation, aiming to enhance the generalization ability of polyp detection models. The dataset consists of 55,883 high-quality endoscopic image frames, including 29,640 polyp frames and 26,243 non-polyp frames. All images have undergone strict screening to eliminate low-quality samples such as blurry ones, images with reflections and ghosting. It features high image quality, balanced class distribution and diverse scenarios, making it suitable for polyp detection, segmentation and data augmentation tasks in medical image analysis, and can be widely applied to the development and evaluation of colonoscopy-assisted diagnosis systems.