Introduction
In order to promote the research of caricature attribute recognition, we construct a face attribute dataset, namely WebCariA, by labeling all images in the WebCaricature dataset. The dataset contains 6024 caricatures and 5974 photos of 252 people. Each image is labeled more than three times with 50 attributes of intrinsic face characters, and the final labels are determined by a voting strategy.
Specifically, to help understand the intrinsic facial characteristics that are felt by human, we did not label the attributes that are changeable, such as 'WearingJewelry', 'WearingHats', etc. Instead, we labeled purely the facial attributes that can reflect the intrinsic characteristics of the faces such as 'BigNose', 'Bald', etc.
Sample Images
The first three column are photos with the attribute and the next three columns are caricatures with that attribute. |
Downloads
-
Contact Jing Huo (huojing@nju.edu.cn) if you want to apply for the dataset.
Citation
The database is made available only for non-commercial use. If you use this dataset, please cite the following paper:
"WebCaricature: A Benchmark for Caricature Recognition"[bibtex] [arXiv] [webpage]
Jing Huo, Wenbin Li, Yinghuan Shi, Yang Gao and Hujun Yin,
British Machine Vision Conference(BMVC), 2018.
"Unsupervised Domain Attention Adaptation Network for Caricature Attribute Recognition"[bibtex] [webpage]
Wen Ji, Kelei He, Jing Huo, Zheng Gu, Yang Gao,
European Conference on Computer Vision(ECCV), 2020.
Copyright
The dataset is released for research and educational purposes. We hold no liability for any undesirable consequences of using the dataset. All rights of the WebCariA Dataset are reserved. Any person or organization is not permitted to distribute, redistribute, publish, copy or disseminate this dataset.