Office + Caltech datasets
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/jindongwang/transferlearning/blob/master/data/dataset.md#office+caltech
下载链接
链接失效反馈官方服务：
资源简介：
该数据集名为Office数据集，是一个视觉领域适应性的基准测试，涵盖了10个常见物体类别，并分为三个不同的子数据集（亚马逊、单反相机和网络摄像头），同时还结合了用于物体识别的标准Caltech数据集。这些子数据集基于提取并量化为800个频段的SURF特征构建而成。规模上，三个子数据集各自包含10个类别。该数据集的任务是领域适应。
The Office Dataset is a benchmark for visual domain adaptation, covering 10 common object categories and comprising three distinct subsets: Amazon, DSLR, and Webcam. It additionally integrates the standard Caltech dataset for object recognition. All three subsets are constructed using SURF features that have been extracted and quantized into 800 frequency bins. Each of the three subsets includes exactly 10 object categories, and the core task of this dataset is domain adaptation.
提供机构：
Publicly available through GitHub
