jianghaojun Goto Github PK

My name is Haojun Jiang(蒋昊峻) [My Google Scholar], a fourth-year Ph.D. student in the Department of Automation at Tsinghua University, advised by Prof. Gao Huang. Before that, I received my B.E. degree in Automation at Tsinghua University. 😄

Beginning in May (2023), I will be shifting my focus to the research of embodied robotic system.

• The World's First Autonomous Echocardiography (TTE) Robotic System (Keywords: Embodied Medical Robot) (Under Progress) [Beijing Municipal Government News][Tsinghua News][The Beijing News]
• Towards Expert-level Autonomous Ultrasonography Using AI-Driven Robotic System (Keywords: Embodied Medical Robot) (Under Progress)
• Sequence-aware Pre-training (Keywords: Vision-Action Pre-training, Sequential modeling, Joint 2D-3D Structure Modeling) (Under Progress) [Paper][Code]
• Structure-aware Cardiac World Model (Keywords: Vision-Action Pre-training, Joint 2D-3D Structure Modeling) (MICCAI'24 5th ASMUS Workshop Oral) [Paper][Code]
• Cardiac Copilot for Echocardiography Probe Guidance (Keywords: Vision-Action Pre-training, 3D Structure Modeling) (MICCAI'24 Early Accept) [Paper][Code]
• Cross-Modal Adapter (Keywords: Vision-Language Alignment, Text-Image/Video Retrieval) (Under Review) [Paper][Code]
• Pseudo-Q (Keywords: Vision-Language Alignment) (CVPR'22) [Paper][Code]
• CondenseNetV2 (Keywords: Edge and Lightweight Model Architecture) (CVPR'21) [Paper][Code]
• AdaFocus (Keywords: Efficicent and Dynamic Video Recognition Framework) (ICCV'21) [Paper][Code]

• Awesome Parameter Efficient Transfer Learning [Repo]
• Awesome Autonomous Medical Ultrasound System [Repo]

[2024/07]: Structure-aware World Model for Probe Guidance via Large-scale Self-supervised Pre-train is accepted by MICCAI 2024 5th ASMUS Workshop and is available at arXiv!
[2024/06]: We were selected into the top ten student innovation and entrepreneurship teams of Tsinghua University - Tsinghua University's highest honor for innovation and entrepreneurship!
[2024/06]: Cardiac Copilot: Automatic Probe Guidance for Echocardiography with World Model now is available at arXiv!
[2024/06]: Our intelligent autonomous ultrasound robot won the Gold Award at the Beijing Challenge Cup Entrepreneurship Competition and was recommended to represent Tsinghua University in national competitions.
[2024/06]: I was selected for Tsinghua University iCenter's Maker Star Program.
[2024/05]: Cardiac Copilot: Automatic Probe Guidance for Echocardiography with World Model is early accepted by MICCAI 2024 (Top 11%)!
[2024/04]: Our intelligent autonomous ultrasound robot won the Silver Award at the Tsinghua Challenge Cup Entrepreneurship Competition and was recommended to represent Tsinghua University in Beijing municipal competitions.
[2024/03]: I was selected for the Tsinghua University 'Qi Chuang' Student Entrepreneurship Talent Program (20/60000) due to our innovative prototype in the field of intelligent autonomous ultrasound robots.
[2024/01]: Our intelligent autonomous ultrasound robot won the First Prize (1/245) at the Global Artificial Intelligence and Robotics Innovation Competition organized by the Guoqiang Research Institute of Tsinghua University.
[2023/12]: Our intelligent autonomous ultrasound robot has been selected as one of the Top-10 in the Tsinghua Medical-Engineering Innovation Competition.
[2023/07]: Deep Incubation: Training Large Models by Divide-and-Conquering is accepted by ICCV 2023! Paper is available at arXiv.
[2023/05]: I will be shifting my focus to the research of autonomous medical ultrasound systems.
[2023/01]: Text4Point now is available at arXiv. This work propose a novel Text4Point framework to construct language-guided 3D point cloud models. The key idea is utilizing 2D images as a bridge to connect the point cloud and the language modalities.
[2022/12]: A curated list about Parameter Efficient Transfer Learning in computer vision and multimodal is created.
[2022/12]: Deep Incubation: Training Large Models by Divide-and-Conquering now is available at arXiv. This work explores a novel Modular Training paradigm which divides a large model into smaller modules, trains them independently, and reassembles the trained modules to obtain the target model.
[2022/11]: Cross-Modal Adapter now is available at arXiv. This work explores the adapter-based parameter-efficient transfer learning for text-video retrieval domain. It reduces 99.6% of fine-tuned parameters without performance degradation.
[2022/09]: An introduction about Parameter Efficient Transfer Learning is given in BAAI dynamic neural network seminar.
[2022/07]: Glance and Focus Networks for Dynamic Visual Recognition is accepted by TPAMI (IF=24.31)!
[2022/07]: AI Time invites me to give a talk about Pseudo-Q.
[2022/04]: An introduction about 3D Visual Grounding is given in BAAI dynamic neural network seminar.
[2022/04]: A curated list about 3D Vision and Language is created.
[2022/03]: Pseudo-Q and AdaFocusV2 are accepted by CVPR 2022!
[2021/07]: AdaFocus is accepted by ICCV 2021!
[2021/03]: CondenseNetV2 is accepted by CVPR 2021!

• Conference Reviewer: CVPR, ICCV, ECCV

Please include a brief note about the reason for reaching out when you contact me.

• E-mail：jianghaojunthu at 163.com
• Wechat：LebronJames5Champ