Synthetic Multi-class Surgical Datasets
收藏资源简介:
合成多类手术数据集是由德国德累斯顿大学医院卡尔古斯塔夫卡鲁斯分校医学院的研究团队创建的,旨在通过扩散模型生成高质量的手术图像及其标注。数据集的内容包括多种手术场景下的器官图像及其对应的语义分割掩码。创建过程涉及使用真实手术图像和分割掩码训练扩散模型,并通过图像合成步骤确保结构和纹理的一致性。该数据集主要应用于计算机辅助手术中的器官识别和分割任务,旨在提高手术场景的理解和术中辅助。
Synthetic Multi-Class Surgical Dataset was created by a research team from the Faculty of Medicine, Carl Gustav Carus University Hospital, Technische Universität Dresden, Germany. It aims to generate high-quality surgical images and their annotations using diffusion models. The dataset includes organ images under various surgical scenarios and their corresponding semantic segmentation masks. The dataset creation process involves training diffusion models with real surgical images and segmentation masks, and ensuring structural and texture consistency through image synthesis steps. This dataset is primarily applied to organ recognition and segmentation tasks in computer-assisted surgery, with the goal of enhancing surgical scene understanding and intraoperative assistance.

- 1Synthesizing Multi-Class Surgical Datasets with Anatomy-Aware Diffusion Models德国德累斯顿大学医院卡尔古斯塔夫卡鲁斯分校医学院 · 2024年



