Jiaqi Liu

I am a second-year Ph.D. student at the University of Trento, supervised by Prof. Nicu Sebe. My current research interests are in generative models for visual applications.

Before my Ph.D. study, I got both of my master and bachelor degrees from Shanghai Jiao Tong University, majored in automation.

Email  /  CV  /  Scholar  /  Github

profile photo

Publications

Multi-focal Conditioned Latent Diffusion for Person Image Synthesis
Jiaqi Liu, Jichao Zhang, Paolo Rota, Nicu Sebe
CVPR, 2025
code / arXiv

Proposes multi-focal condtions to refine the deteriorated sensitive regions for person image synthesis.

Visual-Kinematics Graph Learning for Procedure-Agnostic Instrument Tip Segmentation in Robotic Surgeries.
Jiaqi Liu, Yonghao Long, Kai Chen, Cheuk Hei Leung, Zerui Wang, Qi Dou
IROS, 2023
paper / arXiv

First work to aggregate kinematic data to improve the procedure-agnostic robotic instrument segmentation by graph learning.

Discriminative Asymmetric Learning for Efficient Surgical Instrument Parsing
Jiaqi Liu, Yu Qiao, Jie Yang, Guang-zhong Yang, Yun Gu,
ICRA, 2021
paper

Discriminative learning in dual-branch network to efficiently parse the details and semantic target in real-time robotic parsing.

Work Experience

  • Chinese University of Hong Kong, Hong Kong. 10/2022-7/2023.
    Research Assistant. focusing on kinematic guided robotic parsing.
    Supervisor: Prof. Qi Dou.

  • Citic Securities, Shanghai, China. 11/2021-12/2021.
    Research Internship, focusing on deep-learning based asset allocation and quantitative investment.

  • Robotic X, Tencent Inc, Shenzhen, China. 04/2021-09/2021.
    Research Internship, focusing on grasping pose estimation.
    Mentor: Dr. Yonggen Ling.

Awards

  • Excellent Graduate Award, SJTU, 2022
  • National Scholarship, 2021
  • Excellent Graduate Award, SJTU, 2019
  • Grand Prize, 5th Delta Advanced Automation Contest, 2018

  • Many thanks to Jon Barron of the webpage codes.