CUFSF

该数据集名为CUFSF，包含了1194对人脸照片与素描的配对。所有照片都是在不同的光照条件下获取的。为了处理方便，这些图像都被裁剪并重新调整了形状，但需要注意的是，非对齐和低保真的样本可能会影响最终的性能表现。该数据集的规模为1194对，所涉及的任务是人脸素描合成。

The dataset named CUFSF contains 1194 matched pairs of facial photographs and sketches. All photos were captured under varying lighting conditions. For processing convenience, these images have been cropped and resized. However, it should be noted that misaligned and low-fidelity samples may negatively affect the final performance. The core task associated with this dataset is facial sketch synthesis.

搜集汇总

背景与挑战

背景概述

CUHK Face Sketch FERET Database (CUFSF)是一个用于人脸素描合成与识别研究的数据集，包含1,194个人的数据，每个人对应一张带光照变化的面部照片和一张由艺术家绘制的形状夸张的素描。该数据集基于FERET数据库构建，适用于计算机视觉领域的面部图像处理任务。