SketchRef
收藏资源简介:
SketchRef数据集由北京大学王选计算机研究所创建,包含四个子数据集:动物、人类面部、人类身体和常见物体。该数据集旨在评估自动草图合成算法的质量,特别关注结构识别性和简化级别。数据集的创建过程包括收集参考照片、合成草图以及共享草图和参考照片之间的注释。SketchRef数据集的应用领域主要集中在图像合成领域,特别是草图合成算法的评估和改进。
The SketchRef dataset was developed by the Wangxuan Institute of Computer Technology, Peking University. It comprises four sub-datasets: animals, human faces, human bodies, and common objects. This dataset is designed to evaluate the quality of automated sketch synthesis algorithms, with particular focus on structural recognizability and simplification level. The dataset construction process includes collecting reference photographs, synthesizing sketches, and creating annotations that establish the correspondence between each sketch and its corresponding reference photograph. The main application scenarios of the SketchRef dataset lie in the field of image synthesis, especially for the evaluation and improvement of sketch synthesis algorithms.




