LSUN

LSUN是由普林斯顿大学构建的大型图像数据集，包含约100万张标记图像，涵盖10个场景类别和20个对象类别。该数据集通过深度学习与人工标注相结合的方式创建，旨在解决现有数据集规模和密度不足的问题。LSUN数据集的创建过程涉及从大量候选图像中迭代采样，人工标注，以及使用训练模型分类，最终形成高质量的标注数据。该数据集主要应用于视觉识别领域，支持深度学习模型训练，以提升图像分类和识别的性能。

LSUN is a large-scale image dataset constructed by Princeton University. It contains approximately 1 million labeled images covering 10 scene categories and 20 object categories. Developed by combining deep learning techniques and manual annotation, this dataset aims to address the shortcomings of existing datasets in terms of scale and density. The construction process of LSUN involves iterative sampling from a large pool of candidate images, manual annotation, and classification using trained models, ultimately producing high-quality labeled data. This dataset is primarily utilized in the field of visual recognition, supporting the training of deep learning models to enhance the performance of image classification and recognition.