Endoscopic Videos Dataset

该数据集包含了来自21位患者的临床内窥镜视频，用于评估内窥镜图像匹配技术。这些视频以灰度图像形式呈现，并采用CLAHE方法进行了增强处理，关键点则是通过手工制作的方法检测出来。为了创建图表视图，数据集还对图像进行了旋转、平移和缩放等变换。在交叉验证方案中，该数据集包括21位患者的视频序列，其中16个序列用于训练，5个序列用于验证。该数据集的任务是评估内窥镜图像匹配及特征描述符的效果。

This dataset comprises clinical endoscopic videos from 21 patients, which is designed for evaluating endoscopic image matching techniques. All videos are converted to grayscale images and enhanced using the CLAHE (Contrast Limited Adaptive Histogram Equalization) algorithm. Keypoints in these images are detected via handcrafted methods. To generate various graphical views, the dataset also applies transformations including rotation, translation and scaling to the images. For the cross-validation scheme, the dataset includes video sequences from the 21 patients, with 16 sequences allocated for training and 5 for validation. The primary task of this dataset is to evaluate the performance of endoscopic image matching approaches and feature descriptors.