Changqian Yu

As a Senior Research Scientist at Kunlun Tech, I lead a team developing cutting-edge large multimodal models, with expertise in Diffusion and VLM technologies. My tenure at Meituan’s Autonomous Delivery Department as a Research Scientist equipped me with a robust understanding of real-world applications of AI.

I obtained my Ph.D. degree from School of Artificial Intelligence and Automation, Huazhong University of Science and Technology in 2021, under the expert guidance of Prof. Nong Sang and Prof. Changxin Gao. My academic journey was further enriched by a visiting Ph.D. position at the Australian Institute for Machine Learning, University of Adelaide, where I worked closely with Prof. Chunhua Shen. During my doctoral studies, I received the Excellent Doctoral Dissertation Award (one of only 10 students in the country).

My industry experience includes being a research intern at Microsoft Research Asia and Megvii (Face++), where I collaborated with Dr. Jingdong Wang, Dr. Gang Yu, and Dr. Jian Sun. I was part of the winning team at the COCO & Mapillary Panoptic Segmentation Challenge 2018.

My passion lies in tackling complex problems in computer vision and artificial intelligence. I have made significant contributions to the fields of semantic/panoptic segmentation, behavior prediction, and vision-language models. Among my most notable works are:

BiSeNet (Rank 10 in ECCV 2018 Top-10 Influential Papers & Citation > 2500).
BiSeNetV2 (IJCV & ESI & ESI-Hot & Citation > 1000).
DFNet (CVPR & Citation > 900)
Lite-HRNet (CVPR & Citation > 300)

News

Oct 04, 2023	Ranked as the World’s Top 2% most-cited scientists 2023 by Stanford University.
Nov 23, 2022	Delighted to receive CSIG Excellent Doctoral Dissertation Award/中国图象图形学学会优秀博士学位论文奖 {Media Reports}.
Sep 03, 2022	The IJCV Paper of BiSeNetV2 has been selected as the ESI Paper and ESI-Hot Paper🔥
Aug 15, 2022	Thank Bilibili Vlogger “在下小苏” for the nice interview. {Bilibili Video 1} {Bilibili Video 2} {Youtube Video}.
Jun 15, 2021	Successfully defended my Ph.D. thesis!
Sep 26, 2018	As a team member of Megvii(Face++) and R4D Team, we won the 1st place of COCO Panoptic Segmentation and Mapillary Panoptic Segmentation in the COCO & Mapillary Panoptic Segmentation Challenge 2018. I was invited to present at COCO & Mapillary Joint Workshop at ECCV2018 in Munich, Germany.

Selected Publications

ECCV

BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation

Changqian Yu^*, Jingbo Wang^*, Chao Peng, Changxin Gao^†, Gang Yu, and Nong Sang

In European Conference on Computer Vision (ECCV) , 2018

ECCV 2018 Top-10 Influential Papers arXiv Bib HTML Code

Rank 10 in ECCV 2018 Top-10 Influential Papers

@inproceedings{yu2018bisenet,
  title = {BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation},
  author = {Yu, Changqian and Wang, Jingbo and Peng, Chao and Gao, Changxin and Yu, Gang and Sang, Nong},
  booktitle = {European Conference on Computer Vision},
  pages = {325--341},
  year = {2018},
  google_scholar_id = {UeHWp8X0CEIC},
}

IJCV

BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation

Changqian Yu, Changxin Gao^†, Jingbo Wang, Gang Yu, Chunhua Shen, and Nong Sang

International Journal of Computer Vision (IJCV), 2021

ESI-Hot arXiv Bib HTML Code

ESI-Hot Paper & ESI-Highly Cited Paper

@article{yu2021bisenet,
  title = {BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation},
  author = {Yu, Changqian and Gao, Changxin and Wang, Jingbo and Yu, Gang and Shen, Chunhua and Sang, Nong},
  journal = {International Journal of Computer Vision},
  volume = {129},
  pages = {3051--3068},
  year = {2021},
  google_scholar_id = {Y0pCki6q_DkC},
}

CVPR

Learning a Discriminative Feature Network for Semantic Segmentation

Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, and Nong Sang

In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2018

arXiv Bib HTML Code

@inproceedings{yu2018learning,
  title = {Learning a Discriminative Feature Network for Semantic Segmentation},
  author = {Yu, Changqian and Wang, Jingbo and Peng, Chao and Gao, Changxin and Yu, Gang and Sang, Nong},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
  pages = {1857--1866},
  year = {2018},
  google_scholar_id = {u5HHmVD_uO8C},
}

CVPR

Lite-HRNet: A Lightweight High-Resolution Network

Changqian Yu, Bin Xiao, Changxin Gao, Lu Yuan, Lei Zhang, Nong Sang, and Jingdong Wang

In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2021

arXiv Bib HTML Code

@inproceedings{yu2021lite,
  title = {Lite-HRNet: A Lightweight High-Resolution Network},
  author = {Yu, Changqian and Xiao, Bin and Gao, Changxin and Yuan, Lu and Zhang, Lei and Sang, Nong and Wang, Jingdong},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
  pages = {10440--10450},
  year = {2021},
  google_scholar_id = {WF5omc3nYNoC},
}

CVPR

Context prior for scene segmentation

Changqian Yu, Jingbo Wang, Changxin Gao, Gang Yu, Chunhua Shen, and Nong Sang

In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2020

arXiv Bib HTML Code

@inproceedings{yu2020context,
  title = {Context prior for scene segmentation},
  author = {Yu, Changqian and Wang, Jingbo and Gao, Changxin and Yu, Gang and Shen, Chunhua and Sang, Nong},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
  pages = {12416--12425},
  year = {2020},
  google_scholar_id = {Tyk-4Ss8FVUC},
}

ECCV

Representative Graph Neural Network

Changqian Yu, Yifan Liu, Changxin Gao, Chunhua Shen, and Nong Sang

In European Conference on Computer Vision (ECCV) , 2020

arXiv Bib HTML

@inproceedings{yu2020representative,
  title = {Representative Graph Neural Network},
  author = {Yu, Changqian and Liu, Yifan and Gao, Changxin and Shen, Chunhua and Sang, Nong},
  booktitle = {European Conference on Computer Vision},
  pages = {379--396},
  year = {2020},
  organization = {Springer},
  google_scholar_id = {W7OEmFMy1HYC},
}

IEEE SPL

CondNet: Conditional Classifier for Scene Segmentation

Changqian Yu, Yuanjie Shao, Changxin Gao, and Nong Sang

IEEE Signal Processing Letters (IEEE SPL), 2021

arXiv Bib HTML Code

@article{yu2021condnet,
  title = {CondNet: Conditional Classifier for Scene Segmentation},
  author = {Yu, Changqian and Shao, Yuanjie and Gao, Changxin and Sang, Nong},
  journal = {IEEE Signal Processing Letters},
  volume = {28},
  pages = {758--762},
  year = {2021},
  publisher = {IEEE},
  google_scholar_id = {ufrVoPGSRksC},
}

IEEE TMM

Hybrid motion representation learning for prediction from raw sensor data

Depu Meng, Changqian Yu^†, Jiajun Deng, Deheng Qian, Houqiang Li, and Dongchun Ren

IEEE Transactions on Multimedia (IEEE TMM), 2023

Bib HTML

@article{meng2023hybrid,
  title = {Hybrid motion representation learning for prediction from raw sensor data},
  author = {Meng, Depu and Yu, Changqian and Deng, Jiajun and Qian, Deheng and Li, Houqiang and Ren, Dongchun},
  journal = {IEEE Transactions on Multimedia},
  volume = {25},
  pages = {8868--8879},
  year = {2023},
  publisher = {IEEE},
  google_scholar_id = {MXK_kJrjxJIC},
}

NeurIPS

PLIP: Language-Image Pre-training for Person Representation Learning

Jialong Zuo, Jiahao Hong, Feng Zhang, Changqian Yu, Hanyu Zhou, Changxin Gao, Nong Sang, and Jingdong Wang

In Annual Conference on Neural Information Processing Systems (NeurIPS) , 2024

arXiv Bib Code

@inproceedings{zuo2023plip,
  title = {PLIP: Language-Image Pre-training for Person Representation Learning},
  author = {Zuo, Jialong and Hong, Jiahao and Zhang, Feng and Yu, Changqian and Zhou, Hanyu and Gao, Changxin and Sang, Nong and Wang, Jingdong},
  booktitle = {Annual Conference on Neural Information Processing Systems},
  year = {2024},
  google_scholar_id = {3fE2CSJIrl8C},
}

arXiv

FLUX that Plays Music

Zhengcong Fei, Mingyuan Fan, Changqian Yu, and Junshi Huang

arXiv preprint (arXiv), 2024

arXiv Bib Code

@article{fei2024flux,
  title = {FLUX that Plays Music},
  author = {Fei, Zhengcong and Fan, Mingyuan and Yu, Changqian and Huang, Junshi},
  journal = {arXiv preprint},
  year = {2024},
  google_scholar_id = {M3ejUd6NZC8C},
}

arXiv

Scaling Diffusion Transformers to 16 Billion Parameters

Zhengcong Fei, Mingyuan Fan, Changqian Yu, Debang Li, and Junshi Huang

arXiv preprint (arXiv), 2024

arXiv Bib Code

@article{fei2024scaling,
  title = {Scaling Diffusion Transformers to 16 Billion Parameters},
  author = {Fei, Zhengcong and Fan, Mingyuan and Yu, Changqian and Li, Debang and Huang, Junshi},
  journal = {arXiv preprint},
  year = {2024},
  google_scholar_id = {4TOpqqG69KYC},
}

arXiv

Dimba: Transformer-Mamba Diffusion Models

Zhengcong Fei, Mingyuan Fan, Changqian Yu, Debang Li, Youqiang Zhang, and Junshi Huang

arXiv preprint (arXiv), 2024

arXiv Bib HTML Code

@article{fei2024dimba,
  title = {Dimba: Transformer-Mamba Diffusion Models},
  author = {Fei, Zhengcong and Fan, Mingyuan and Yu, Changqian and Li, Debang and Zhang, Youqiang and Huang, Junshi},
  journal = {arXiv preprint},
  year = {2024},
  google_scholar_id = {YOwf2qJgpHMC},
}

arXiv

Scalable diffusion models with state space backbone

Zhengcong Fei, Mingyuan Fan, Changqian Yu, and Junshi Huang

arXiv preprint (arXiv), 2024

arXiv Bib Code

@article{fei2024scalable,
  title = {Scalable diffusion models with state space backbone},
  author = {Fei, Zhengcong and Fan, Mingyuan and Yu, Changqian and Huang, Junshi},
  journal = {arXiv preprint},
  year = {2024},
  google_scholar_id = {Zph67rFs4hoC},
}

IEEE TIP

Conditional boundary loss for semantic segmentation

Dongyue Wu, Zilin Guo, Aoyan Li, Changqian Yu, Changxin Gao, and Nong Sang

IEEE Transactions on Image Processing (IEEE TIP), 2023

Bib HTML Code

@article{wu2023conditional,
  title = {Conditional boundary loss for semantic segmentation},
  author = {Wu, Dongyue and Guo, Zilin and Li, Aoyan and Yu, Changqian and Gao, Changxin and Sang, Nong},
  journal = {IEEE Transactions on Image Processing},
  year = {2023},
  publisher = {IEEE},
  google_scholar_id = {kNdYIx-mwKoC},
}

CVPR

An End-to-End Network for Panoptic Segmentation

Huanyu Liu, Chao Peng, Changqian Yu, Jingbo Wang, Xu Liu, Gang Yu, and Wei Jiang

In IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2019

arXiv Bib HTML

@inproceedings{liu2019end,
  title = {An End-to-End Network for Panoptic Segmentation},
  author = {Liu, Huanyu and Peng, Chao and Yu, Changqian and Wang, Jingbo and Liu, Xu and Yu, Gang and Jiang, Wei},
  booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},
  pages = {6172--6181},
  year = {2019},
  google_scholar_id = {2osOgNQ5qMEC},
}

ECCV

Efficient Semantic Video Segmentation with Per-frame Inference

Yifan Liu, Chunhua Shen, Changqian Yu, and Jingdong Wang

In European Conference on Computer Vision (ECCV) , 2020

arXiv Bib HTML Code

@inproceedings{liu2020efficient,
  title = {Efficient Semantic Video Segmentation with Per-frame Inference},
  author = {Liu, Yifan and Shen, Chunhua and Yu, Changqian and Wang, Jingdong},
  booktitle = {European Conference on Computer Vision},
  pages = {352--368},
  year = {2020},
  organization = {Springer},
  google_scholar_id = {zYLM7Y9cAGgC},
}