Huiyu LI

Master @ BIT

PhD @ INRIA

Email: huiyu.li@inria.fr

Huiyu LI

Hello! I am actively seeking an exciting and challenging postdoctoral position.

I received my Ph.D. from INRIA in Sophia Antipolis, France, under the guidance of Prof. Hervé Delingette and Prof. Nicholas Ayache. I was also fortunate to receive support from the French government through the 3IA Côte d’Azur investments program. Previously, I completed a master's degree in Computer Science at Beijing Institute of Technology.

The goal of my research is to empower AI to advance healthcare. I am particularly interested in applying generative models to medical image anonymization, enabling the removal of patient-identifiable information while preserving diagnostic information. This challenging task involves achieving a delicate balance between image reconstruction and anonymization, as well as identity removal and utility preservation. Outside of my academic pursuits, I enjoy writing, gardening, skiing, and playing badmiminton.

Dear Like-minded Friends, If something I share resonates with you or touches your heart, I'd love to hear from you! Feel free to reach out via email or schedule a time to chat through Chatting Time. Let's connect and see the world through each other's eyes. I've always believed that everyone is like a book. My pages are filled with fashion design, architecture, interior decoration, gardening, cooking, drawing, and dreaming. I look forward to sharing and learning together!

News:

Nov 2024: I denfend my PhD thesis, Data Exfiltration and Anonymization of Medical Images based on Generative Models.

Generative Medical Image Anonymization Based on Latent Code Projection and Optimization

Huiyu-Li, Nicholas Ayache, Hervé Delingette

International Symposium on Biomedical Imaging (ISBI), 2025

arXiv

code

                                @article{li2025generative,
                                    title={Generative medical image anonymization based on latent code projection and optimization},
                                    author={Li, Huiyu and Ayache, Nicholas and Delingette, Herv{\'e}},
                                    journal={arXiv preprint arXiv:2501.09114},
                                    year={2025}
                                  }

Data Stealing Attack on Medical Images: Is it Safe to Export Networks from Data Lakes ?

Huiyu-Li, Nicholas Ayache, Hervé Delingette*

Distributed, Collaborative, and Federated Learning, and Affordable AI and Healthcare for Resource Diverse Global Health. DeCaF FAIR, 2022

arXiv

                                @inproceedings{li2022data,
                                    title={Data Stealing Attack on Medical Images: Is it Safe to Export Networks from Data Lakes?},
                                    author={Li, Huiyu and Ayache, Nicholas and Delingette, Herv{\'e}},
                                    booktitle={International Workshop on Distributed, Collaborative, and Federated Learning},
                                    pages={28--36},
                                    year={2022},
                                    organization={Springer}
                                    }

Deep Distance Map Regression Network with Shape-aware Loss for Imbalanced Medical Image Segmentation

Huiyu LI, Xiabi Liu, Said Boumaraf, Xiaopeng Gong, Donghai Liao, Xiaohong Ma

International Workshop on Machine Learning in Medical Imaging, 2020

Code

                                @inproceedings{li2020deep,
                                    title={Deep distance map regression network with shape-aware loss for imbalanced medical image segmentation},
                                    author={Li, Huiyu and Liu, Xiabi and Boumaraf, Said and Gong, Xiaopeng and Liao, Donghai and Ma, Xiaohong},
                                    booktitle={Machine Learning in Medical Imaging: 11th International Workshop, MLMI 2020, Held in Conjunction with MICCAI 2020, Lima, Peru, October 4, 2020, Proceedings 11},
                                    pages={231--240},
                                    year={2020},
                                    organization={Springer}
                                  }

A New Three-stage Curriculum Learning Approach to Deep Network based Liver Tumor Segmentation

Huiyu LI, Xiabi Liu, Said Boumaraf, Weihua Liu, Xiaopeng Gong, Xiaohong Ma

International Joint Conference on Neural Networks (IJCNN), 2020

arXiv

code

                                @inproceedings{li2020new,
                                    title={A new three-stage curriculum learning approach for deep network based liver tumor segmentation},
                                    author={Li, Huiyu and Liu, Xiabi and Boumaraf, Said and Liu, Weihua and Gong, Xiaopeng and Ma, Xiaohong},
                                    booktitle={2020 International Joint Conference on Neural Networks (IJCNN)},
                                    pages={1--6},
                                    year={2020},
                                    organization={IEEE}
                                  }

A Dynamic Scale-Free Network Particle Swarm Optimization for Extracting Features on Multi-Omics Data

Huiyu LI, Sheng-Jun Li, Junliang Shang*, Jin-Xing Liu, Chun-Hou Zheng

Journal of Computational Biology, 2018

paper

code

                                @article{li2019dynamic,
                                    title={A dynamic scale-free network particle swarm optimization for extracting features on multi-omics data},
                                    author={Li, Huiyu and Li, Sheng-Jun and Shang, Junliang and Liu, Jin-Xing and Zheng, Chun-Hou},
                                    journal={Journal of Computational Biology},
                                    volume={26},
                                    number={8},
                                    pages={769--781},
                                    year={2019},
                                    publisher={Mary Ann Liebert, Inc., publishers 140 Huguenot Street, 3rd Floor New~…}
                                  }

An Improved Particle Swarm Optimization with Dynamic Scale-Free Network for Detecting Multi-Omics Features

Huiyu LI, Sheng-Jun Li*, Junliang Shang, Jin-Xing Liu, Chun-Hou Zheng

International Symposium on Bioinformatics Research and Applications, 2018

paper

                                @inproceedings{li2018improved,
                                    title={An improved particle swarm optimization with dynamic scale-free network for detecting multi-omics features},
                                    author={Li, Huiyu and Li, Sheng-Jun and Shang, Junliang and Liu, Jin-Xing and Zheng, Chun-Hou},
                                    booktitle={Bioinformatics Research and Applications: 14th International Symposium, ISBRA 2018, Beijing, China, June 8-11, 2018, Proceedings 14},
                                    pages={26--37},
                                    year={2018},
                                    organization={Springer}
                                  }

Beyond my research, I am a writer. I write not for sale, but for my soul—especially when I cannot find a book that truly entertains it.

As the saying goes, words come from life. In that sense, I see myself as a “generator,” learning from real-life experiences and transforming them into written words. I am not a very good liar, nor am I a model that can easily “unlearn”. My words inevitably reflect my personality and my life, making it easy to trace who I am. For this reason, much of my writing may never be publicly published while I am alive. However, I am always happy to share the core thoughts and inspirations behind my words.

I have always believed that everyone is like a book, worth reading and appreciating.

Three Posonalities

This explores the idea that everyone has three personalities, shaped by ideals, reality, and society. The ideal self represents who we truly aspire to be, free from societal constraints. The social self conforms to societal norms and expectations. For most people, life is about finding a balance or compromise between these two, ultimately shaping the real self—a blend of personal desires and external influences.

Society

Growth

Personality
Starry Sky

This is inspired by an ancient belief that when our loved ones pass away, they become stars in the sky. The protagonist envisions a virtual parallel space where they can communicate with their lost loved ones in the darkness sky. Over time, as more stars appear, the once-empty sky transforms into a vast, starry expanse—mirroring the protagonist's journey of growth and aging.

Family

Growth
The Substitute

The already married hero unexpectedly finds the heroine as an alternative of his died extra-girlfriend, while the heroine finds the lost love from her young died father from the hero. They try to review the lost lives and the memories with each other as an alternative, and finally encourage each other to accept the real life.

Love

Images in this subsection are credited to DeepAI.

Huiyu LI

PUBLICATIONS

Generative Medical Image Anonymization Based on Latent Code Projection and Optimization

Data Stealing Attack on Medical Images: Is it Safe to Export Networks from Data Lakes ?

Deep Distance Map Regression Network with Shape-aware Loss for Imbalanced Medical Image Segmentation

A New Three-stage Curriculum Learning Approach to Deep Network based Liver Tumor Segmentation

A Dynamic Scale-Free Network Particle Swarm Optimization for Extracting Features on Multi-Omics Data

An Improved Particle Swarm Optimization with Dynamic Scale-Free Network for Detecting Multi-Omics Features

STORIES

Three Posonalities

Starry Sky

The Substitute

CHATTING TIME