ROA-BEV: 2D Region-Oriented Attention for BEV-based 3D Object Detection
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025
Biography
Jiwei Chen is a Ph.D. candidate in Computer and Information Engineering at the School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen, advised by Prof. Rui Huang. He received his M.Eng. in Control Science and Engineering from Nanjing University of Information Science and Technology (NUIST), advised by Prof. Yubao Sun, and his B.Eng. in Automation from NUIST.
His research interests include compressed sensing image reconstruction, 3D object detection for autonomous driving, and vision-language-action (VLA) models for driving. His work has been published in Pattern Recognition, IEEE TIP, IEEE TMM, ECCV, and IROS, and he holds several authorized and published patents. His current work focuses on open-world 3D object detection for open roads, and on detecting abnormal-condition objects and understanding scene semantics in driving scenarios. He is open to academic collaborations and research discussions.
News
- 2026.04Started the industrial project Development of Frontier VLA Algorithms with Freetech, serving as the technical lead for architecture, training, and simulation validation.
- Fall 2026Will serve as a teaching assistant for AIE1903: AI Exploration III, contributing to course design and teaching on autonomous driving algorithms.
- 2025Two papers, ROA-BEV and Self-ToF, were accepted to IROS 2025.
- 2023.01Joined the Shenzhen Key Natural Science Foundation project Intelligent Perception System for Automobiles, working on anomaly recognition of road environment changes; the project concluded in 2025.12.
Education
- 2021.09 – 2023.09, 2024.09 – Present The Chinese University of Hong Kong, Shenzhen · Ph.D. in Computer and Information Engineering Advisor: Prof. Rui Huang | Research: 3D object detection for autonomous driving, autonomous driving VLA models
- 2017.09 – 2020.06 Nanjing University of Information Science and Technology · M.Eng. in Control Science and Engineering Advisor: Prof. Yubao Sun | Research: compressed sensing image reconstruction
- 2012.09 – 2014.06, 2015.09 – 2017.06 Nanjing University of Information Science and Technology · B.Eng. in Automation
Research Interests
Compressed Sensing Image Reconstruction
Deep-network-based CS image reconstruction, including generative adversarial networks, dual-path attention, memory-augmented cascading, and non-local regularization methods.
3D Object Detection for Autonomous Driving
BEV-based 3D object detection, and VLM-based open-world 3D detection for 3D localization and open-vocabulary semantic description of known and unknown objects.
Autonomous Driving VLA Models
Architecture, training, and closed-loop validation of vision-language-action models for driving, covering data annotation loops, training acceleration, simulation-based validation, and reinforcement learning.
Publications
Self-Supervised Enhancement for Depth from a Lightweight ToF Sensor with Monocular Images
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025
Learning Image Compressed Sensing with Sub-Pixel Convolutional Generative Adversarial Network
Pattern Recognition (PR), 2020, 98: 107051
Dual-Path Attention Network for Compressed Sensing Image Reconstruction
IEEE Transactions on Image Processing (TIP), 2020
Learning Memory Augmented Cascading Network for Compressed Sensing of Images
European Conference on Computer Vision (ECCV), 2020
Learning Non-Locally Regularized Compressed Sensing Network with Half-Quadratic Splitting
IEEE Transactions on Multimedia (TMM), 2020
Under Review
Physics-Guided Diffusion with Homogeneous Distance Fields for Hand-Object Interaction Generation
Under review at IEEE Transactions on Human-Machine Systems
Projects
Development of Frontier VLA Algorithms
2026.04 – 2027.02An industrial collaboration with Freetech. As the main technical lead, responsible for architecture design and training of the algorithms, covering data annotation closed loops, training acceleration, simulation-based validation, and reinforcement learning training, and driving deployment with the company while distilling problems for future publications.
Intelligent Perception System for Automobiles
2023.01 – 2025.12Studied anomaly recognition for road environment changes, and implemented the algorithms and visualization, meeting the project's target metrics.
Teaching
- 2021 – 2025, 5 semesters CSC3002: C/C++ Programming · Teaching Assistant Course code walkthroughs, design of an online grading system for student assignments, exam grading, and supervision and grading of student projects.
- Fall 2026 (upcoming) AIE1903: AI Exploration III · Teaching Assistant Introducing the development of autonomous driving algorithms, participating in syllabus and content design, course code walkthroughs, exam grading, and supervision and grading of student projects.
Patents
- Yubao Sun, Hongwei Xu, Qingshan Liu, Jiwei Chen, Yi Chen. A hypergraph convolutional network model and its semi-supervised classification method.CN109492691A
- Yubao Sun, Ying Yang, Jiwei Chen, Qingshan Liu. A non-local-regularized compressed sensing network method for image reconstruction.CN110728728B
- Renlong Hang, Qingshan Liu, Jiwei Chen, Sujuan Wang. A hyperspectral image classification method.CN108182449A
- Yubao Sun, Jiwei Chen, Qingshan Liu, Hongwei Xu. A GAN-based method and system for compressed sensing image reconstruction.CN109544656B
- Yubao Sun, Yi Chen, Qingshan Liu, Jiwei Chen. A hyperspectral image classification method based on a label-constrained elastic-net graph model.CN110399909B
- Yubao Sun, Chunfang Chen, Hongwei Xu, Jiwei Chen. A residual dual-attention deep network for kidney segmentation in CT images.CN110675406A
- Rui Huang, Hualie Jiang, Jiwei Chen, Laiyan Ding. An indoor spatial depth estimation method, device, equipment and storage medium.CN113516698B
- Rui Huang, Jiwei Chen, Junliang Li, Yongquan Chen, Zhenxing Xu. An open-world 3D object detection method and device.CN122368977A
Honors & Awards
- 2019 Graduate Merit Scholarship and Second-Class Academic Scholarship, NUIST
- Third Prize, 16th China Postgraduate Mathematical Contest in Modeling ("Huawei Cup")
- Outstanding Graduate, NUIST, Academic Year 2019–2020
- Merit Student, NUIST, Academic Year 2019–2020
- Second Prize, Innovation Achievements Exhibition for 2020 Graduates, NUIST
- Finalist, Outstanding Master's Thesis Award, 4th Jiangsu Artificial Intelligence Conference, 2021