Home

Awesome

iCartoonFace

License: MIT stars GitHub issues GitHub repo size

Cartoon Face Recognition: A Benchmark Dataset

<img align=left src="figures/illustration.png" alt="illustration" style="zoom:100%;" />

The iCartoonFace project is supported by iQIYI. And this repository provides iCartoonFace dataset and baseline approaches of the following paper.

Dataset Download | Citation | Video Presentation | Paper Arxiv(Pdf) | Project Website

Cartoon Face Detection

<img align=left src="figures\detection.png" alt="detection" style="zoom:150%;" />

The iCartoonFace detection dataset is a large-scale dataset established for cartoon face detection, which contains multiple styles. In the iCartoonFace detection task, the mAP (mean average precision) metric is used to evaluate the performance of the algorithm.

Cartoon Face Recognition

<img align=left src="figures\recognition.png" style="zoom:150%;" />

The iCartoonFace recognition dataset is a large-scale challenging dataset established for cartoon face recognition. The above figure visualizes the statistics of the proposed dataset. In the iCartoonFace recognition task, given a probe photo and a gallery containing at least one photo of the same cartoon character, the algorithm needs to rank-orders all photos in the gallery based on similarity to the probe.

Reference method: insightface or reid-strong-baseline

Dataset

Evaluation

Acknowledgement

We would like to thank Song Shi and his team for the organization of competitions, Yan Fu and He Chen team for the help of data annotations, Chenwei Yang team for providing computing resources, and Lingyun Xiao, Ke Chen, Xiang Xia et al. for developing and designing competition websites.

Citation

If you use the iCartoonFace dataset for your research, please cite our paper as follows.

@inproceedings{zheng2020cartoon,
title={Cartoon Face Recognition: A Benchmark Dataset},
author={Zheng, Yi and Zhao, Yifan and Ren, Mengyuan and Yan, He and Lu, Xiangju and Liu, Junhui and Li, Jia},
booktitle={Proceedings of the 28th ACM International Conference on Multimedia},
pages={2264--2272},
year={2020}
}