Wanhua Li

I am an incoming postdoctoral fellow at Harvard University supervised by Prof. Hanspeter Pfister in Fall 2022. Prior to that, I received my Ph.D. from the Department of Automation at Tsinghua University in 2022.

If you are interested in my research or would like to work with me as an intern at Harvard University, feel free to contact me. Remote collaboration is also welcome!

Email: wanhua016 [AT] gmail [DOT] com
~~li-wh17 [AT] mails [DOT] tsinghua [DOT] edu [DOT] cn~~

CV / Google Scholar / Twitter / GitHub

News

2022-09: One paper on language-guided ordinal regression is accepted by NeurIPS 2022.

2022-07: Two papers on multi-attribute learning and talking head synthesis are accepted by ECCV 2022.

2022-06: One paper on age estimation is accepted by TIP.

2022-04: One paper on image inpainting is accepted by TMM.

2021-10: Our team won the 3rd place in 2021 VIPriors Instance Segmentation Challenge (ICCV 2021).

2021-07: One paper on video inpainting detection is accepted by ICCV 2021.

2021-04: One paper on kinship verification is accepted by TIP.

2021-03: Three papers on uncertainty learning, kinship verification, and face clustering are accepted to CVPR 2021.

2020-07: One paper on social relation recognition is accepted to ECCV 2020.

2020-03: One paper on kinship verification is accepted as oral presentation at ICME 2020.

2019-02: One paper on age estimation is accepted to CVPR 2019.

Selected Publications

	Label2Label: A Language Modeling Framework for Multi-Attribute Learning Wanhua Li, Zhexuan Cao, Jianjiang Feng, Jie Zhou, and Jiwen Lu European Conference on Computer Vision (ECCV), 2022 [Website] [arxiv] [Video] [Code] We propose a language modeling framework named Label2Label to model the complex instance-wise attribute relations, which regards each attribute label as a “word” and recovers the label “sentence” based on the masked one.
	Learning Dynamic Facial Radiance Fields for Few-Shot Talking Head Synthesis Shuai Shen, Wanhua Li, Zheng Zhu, Yueqi Duan, Jie Zhou, and Jiwen Lu European Conference on Computer Vision (ECCV), 2022 [Website] [arxiv] [Video] [Code] We propose dynamic facial radiance fields conditioned on the 3D aware reference image features. The facial field can rapidly generalize to novel identities with only 15s clip.
	MetaAge: Meta-Learning Personalized Age Estimators Wanhua Li, Jiwen Lu, Abudukelimu Wuerkaixi, Jianjiang Feng, and Jie Zhou IEEE Transactions on Image Processing, 2022 [Website] [Paper] [arxiv] [Code] We propose a personalized age estimation method named MetaAge, which learns the mapping from identity information to age estimator parameters.
	Learning Adaptive Patch Generators for Mask-Robust Image Inpainting Hongyi Sun, Wanhua Li, Yueqi Duan, Jie Zhou, and Jiwen Lu IEEE Transactions on Multimedia, 2022 [Paper] [bibtex] We propose a Mask-Robust Inpainting Network (MRIN) to recover the masked areas of an image with a patch-wise inpainting process.
	Reasoning Graph Networks for Kinship Verification: from Star-shaped to Hierarchical Wanhua Li, Jiwen Lu, Abudukelimu Wuerkaixi, Jianjiang Feng, and Jie Zhou IEEE Transactions on Image Processing, 2021 [Paper] [arxiv] [bibtex] We develop a Hierarchical Reasoning Graph Network (H-RGN) to exploit more powerful and flexible capacity for graph-based kinship verification.
	Frequency-Aware Spatiotemporal Transformers for Video Inpainting Detection Bingyao Yu, Wanhua Li, Xiu Li, Jiwen Lu, and Jie Zhou IEEE International Conference on Computer Vision (ICCV), 2021 [Paper] [bibtex] We propose a Frequency-Aware Spatiotemporal Transformer for video inpainting detection, which simultaneously mines the traces of video inpainting from spatial, temporal, and frequency domains.
	Learning Probabilistic Ordinal Embeddings for Uncertainty-Aware Regression Wanhua Li, Xiaoke Huang, Jiwen Lu, Jianjiang Feng, and Jie Zhou IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021 [Website] [arxiv] [Video] [Code] We propose probabilistic ordinal embeddings to empower the present-day regression methods with the ability of uncertainty estimation.
	Meta-Mining Discriminative Samples for Kinship Verification Wanhua Li, Shiwei Wang, Jiwen Lu, Jianjiang Feng, and Jie Zhou IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021 [Website] [arxiv] [Video] [bibtex] A Discriminative Sample Meta-Mining strategy is proposed to mine discriminative information from limited positive pairs and sufficient negative samples for kinship verification.
	Structure-Aware Face Clustering on a Large-Scale Graph with 10^7 Nodes Shuai Shen, Wanhua Li, Zheng Zhu, Guan Huang, Dalong Du, Jiwen Lu, and Jie Zhou IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021 [Website] [arxiv] [Code] [Video] It is the first face clustering method to train on very large-scale graph with 20M nodes, and achieve superior inference results on 12M testing data.
	Graph-Based Social Relation Reasoning Wanhua Li, Yueqi Duan, Jiwen Lu, Jianjiang Feng, and Jie Zhou European Conference on Computer Vision (ECCV), 2020 [Website] [arxiv] [Video] [Code] A simpler, faster, and more accurate method for social relation recognition.
	Graph-based Kinship Reasoning Network Wanhua Li, Yingqiang Zhang, Kangchen Lv, Jiwen Lu, Jianjiang Feng, Jie Zhou IEEE International Conference on Multimedia and Expo (ICME), 2020 Oral Presentation [arXiv] [Video] [bibtex] We considers how to compare and fuse the extracted feature pair to reason about the kin relations with the proposed graph-based kinship reasoning networks.
	BridgeNet: A Continuity-Aware Probabilistic Network for Age Estimation Wanhua Li, Jiwen Lu, Jianjiang Feng, Chunjing Xu, Jie Zhou, Qi Tian IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019 [arXiv] [PDF] [bibtex] We propose BridgeNet for age estimation, which aims to mine the continuous relation between age labels effectively.