Synthetic Feature Datasets for Image Matching

Collection of three datasets of synthetic images designed to train image matching models. The first two datasets, called patch dataset and mosaic images dataset, are for training feature detection models (first matching step). The third dataset, called feature pairs dataset, is for training feature description and matching models (last two steps of the matching pipeline). Each dataset is accompanied by an IPYNB notebook code allowing to read and display the data as well as other information (format, type, number of samples) in order to facilitate its use.

本数据集集合包含三组合成图像数据集，旨在用于训练图像匹配模型。其中前两组数据集分别为补丁数据集（patch dataset）与马赛克图像数据集（mosaic images dataset），用于训练特征检测模型，即匹配流程的第一步。第三组数据集为特征对数据集（feature pairs dataset），用于训练特征描述与匹配模型，对应匹配管线的最后两步。每组数据集均附带一份IPYNB格式的笔记本代码，可用于读取、展示数据，同时提供数据格式、类型、样本数量等相关信息，以方便使用者开展相关操作。