SCARED-C
收藏资源简介:
SCARED-C数据集是由香港中文大学电子工程系和新加坡国立大学生物医学工程系共同创建的,旨在评估内窥镜深度估计模型的鲁棒性。该数据集基于SCARED数据集,包含22,950帧图像,涵盖了16种不同类型的合成图像损坏,每种损坏分为五个严重级别。数据集的创建过程包括对原始图像进行多种合成损坏处理,以模拟真实手术环境中的图像失真。该数据集主要应用于内窥镜手术中的深度感知和模型鲁棒性评估,旨在提高手术精度和患者安全。
The SCARED-C dataset was co-developed by the Department of Electronic Engineering, The Chinese University of Hong Kong, and the Department of Biomedical Engineering, National University of Singapore, with the purpose of evaluating the robustness of endoscopic depth estimation models. Built upon the original SCARED dataset, it contains 22,950 image frames, covering 16 distinct types of synthetic image corruptions, each categorized into five severity levels. The dataset construction process applies various synthetic corruptions to the original images to simulate image distortions in real surgical environments. This dataset is primarily utilized for depth perception in endoscopic surgery and model robustness evaluation, aiming to improve surgical accuracy and patient safety.

- 1Benchmarking Robustness of Endoscopic Depth Estimation with Synthetically Corrupted Data香港中文大学电子工程系,新加坡国立大学生物医学工程系 · 2024年



