Image Patch Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/pirofti/ConditionedTransformLearning
下载链接
链接失效反馈官方服务：
资源简介：
该数据集由从 Barbara、Peppers 和 Lena 等知名测试图像中提取的非重叠8x8像素块组成，并进行了向量化处理。此外，数据集包含了用于在不同噪声水平下进行去噪测试的清晰图像和噪声图像。该数据集的规模为 m = 12288，即连结后的向量化块的大小。该数据集的任务领域涉及图像处理和变换学习。
This dataset comprises non-overlapping 8x8 pixel patches extracted from well-known test images such as Barbara, Peppers, and Lena, and has been vectorized. Additionally, the dataset includes both clean and noisy images for denoising tests under different noise levels. The size of this dataset is m = 12288, which refers to the dimension of the concatenated vectorized patches. The task domain of this dataset involves image processing and transform learning.
提供机构：
Publicly available implementations and datasets from various authors
