SERV-CT

SERV-CT数据集是由UCL的Wellcome/EPSRC介入与外科科学中心开发的，用于验证内窥镜3D重建的CT数据集。该数据集包含来自两个离体猪全躯干尸体的16对立体图像，每对图像均经过手动对齐以匹配立体视图，并计算参考视差。数据集涵盖了多种组织类型和显著的深度变化，非常适合评估各种立体算法。

The SERV-CT dataset was developed by the Wellcome/EPSRC Centre for Interventional and Surgical Sciences at UCL, and it is a CT dataset intended for validating endoscopic 3D reconstruction. This dataset contains 16 pairs of stereo images sourced from two entire ex vivo porcine trunks. Each pair of images has been manually aligned to match the stereo view, and reference disparity has been calculated. The dataset covers a variety of tissue types and significant depth variations, making it highly suitable for evaluating various stereo algorithms.