Siang Chen

Siang Chen

Ph.D. Candidate at Department of Electronic Engineering, Tsinghua University.

Posts by Collection

portfolio

publications

Distribution-aware Low-bit Quantization for 3D Point Cloud Networks

Published in 2022 IEEE International Conference on Visual Communications and Image Processing (VCIP), 2022

Various low-bit quantized methods have been widely exploited and shown decent performance on 2D vision tasks in recent years. Complemented with 2D images, 3D point clouds provide an opportunity to understand the surrounding environ-ment better. However, low-bit quantization methods designed for 2D vision tasks are not readily transferable to 3D point clouds due to the higher dimension of 3D data and the increased proportion of activations. … Read more

Recommended citation:

Hu, D., Chen, S., Yang, H., & Wang, G. (2022, December). Distribution-aware Low-bit Quantization for 3D Point Cloud Networks. In 2022 IEEE International Conference on Visual Communications and Image Processing (VCIP) (pp. 1-5). IEEE. [pdf] [bib]

Efficient Heatmap-Guided 6-Dof Grasp Detection in Cluttered Scenes

Published in IEEE Robotics and Automation Letters, 2023

Fast and robust object grasping in clutter is a crucial component of robotics. Most current works resort to the whole observed point cloud for 6-Dof grasp generation, ignoring the guidance information excavated from global semantics, thus limiting high-quality grasp generation and real-time performance. … Read more

Recommended citation:

Chen, S., Tang, W., Xie, P., Yang, W., & Wang, G. (2023). Efficient heatmap-guided 6-DoF grasp detection in cluttered scenes. IEEE Robotics and Automation Letters. [pdf] [bib]

Part-Guided 3D RL for Sim2Real Articulated Object Manipulation

Published in IEEE Robotics and Automation Letters, 2023

Manipulating unseen articulated objects through visual feedback is a critical but challenging task for real robots. Existing learning-based solutions mainly focus on visual affordance learning or other pre-trained visual models to guide manipulation policies, which face challenges for novel instances in real-world scenarios. … Read more

Recommended citation:

Xie, P., Chen, R., Chen, S., Qin, Y., Xiang, F., Sun, T., ... & Su, H. (2023). Part-Guided 3D RL for Sim2Real Articulated Object Manipulation. IEEE Robotics and Automation Letters. [pdf] [bib]

Query-guided Support Prototypes for Few-shot 3D Indoor Segmentation

Published in IEEE Transactions on Circuits and Systems for Video Technology, 2023

Few-shot 3D point cloud segmentation segments novel categories in point cloud scenes with only limited annotations. However, most current methods do not consider query content when exploring support prototypes, and thus suffer from intra-class variations between objects and incomplete representation of category information from annotated support samples. … Read more

Recommended citation:

Hu, D., Chen, S., Yang, H., & Wang, G. (2023). Query-guided Support Prototypes for Few-shot 3D Indoor Segmentation. IEEE Transactions on Circuits and Systems for Video Technology. [pdf] [bib]

Uncertainty-Aware Laser Stripe Segmentation with Non-Local Mechanisms for Welding Robots

Published in under review, 2024

The line-structured-light system has been widely applied in intelligent welding robots for weld seam reconstruction and tracking. However, it’s challenging to extract the projected laser stripes from captured images due to the strong noise and high dynamic range in welding environments. … Read more

Recommended citation: Yixiang Dai, Siang Chen, Tianyu Sun, Zimo Fan, Chun Zhang, Xiaobing Feng, Guijin Wang. (2024). Uncertainty-Aware Laser Stripe Segmentation with Non-Local Mechanisms for Welding Robots.

Rethinking 6-Dof Grasp Detection: A Flexible Framework for High-Quality Grasping

Published in arXiv preprint, 2024

Robotic grasping is a primitive skill for complex tasks and is fundamental to intelligence. For general 6-Dof grasping, most previous methods directly extract scene-level semantic or geometric information, while few of them consider the suitability for various downstream applications, such as target-oriented grasping. … Read more

Recommended citation:

Pengwei Xie, Siang Chen, Wei Tang, Dingchang Hu, Wenming Yang, Guijin Wang. (2024). Rethinking 6-Dof Grasp Detection: A Flexible Framework for High-Quality Grasping. arXiv preprint arXiv:2403.15054. [pdf]

Region-Centric 6-Dof Grasp Detection: A Data-Efficient Solution for Cluttered Scenes

Published in under revision, 2024

Robotic grasping, serving as the cornerstone for complex manipulation tasks, is fundamental for embodied intelligence. For general 6-Dof grasping, most data-driven methods directly extract scene-level information, resorting to fitting with a large amount of data. … Read more

Recommended citation: Wei Tang, Siang Chen, Pengwei Xie, Dingchang Hu, Wenming Yang, Guijin Wang. (2024). Region-Centric 6-Dof Grasp Detection: A Data-Efficient Solution for Cluttered Scenes.

Category-Agnostic Pose Estimation for Point Clouds

Published in IEEE International Conference on Image Processing 2024 (ICIP 2024), 2024

The goal of object pose estimation is to visually determine the pose of a specific object in the RGB-D input. Unfortunately, when faced with new categories, both instance-based and category-based methods are unable to deal with unseen objects of unseen categories, which is a challenge for pose estimation. … Read more

Recommended citation:

Bowen Liu, Wei Liu, Siang Chen, Pengwei Xie, Guijin Wang. (2024). Category-Agnostic Pose Estimation for Point Clouds. [pdf]

Variation-robust Few-shot 3D Affordance Segmentation for Robotic Manipulation

Published in under review, 2024

Traditional affordance segmentation on 3D point cloud objects requires massive amounts of annotated training data and can only make predictions within predefined classes and affordance tasks. … Read more

Recommended citation: Dingchang Hu, Tianyu Sun, Pengwei Xie, Siang Chen, Yixiang Dai, Huazhong Yang, Guijin Wang. (2024). Variation-robust Few-shot 3D Affordance Segmentation for Robotic Manipulation.

Target-Oriented Object Grasping via Multimodal Human Guidance

Published in ECCV 2024 Workshop on Assistive Computer Vision and Robotic (ACVR 2024), 2024

In the context of human-robot interaction and collaboration scenarios, robotic grasping still encounters numerous challenges. Traditional grasp detection methods generally analyze the entire scene to predict grasps, leading to redundancy and inefficiency. … Read more

Recommended citation:

Pengwei Xie, Siang Chen, Dingchang Hu, Yixiang Dai, Kaiqin Yang, Guijin Wang. (2024). Target-Oriented Object Grasping via Multimodal Human Guidance. [pdf]

Region-aware Grasp Framework with Normalized Grasp Space for Efficient 6-DoF Grasping

Published in CoRL 2024, 2024

A series of region-based methods succeed in extracting regional features and enhancing grasp detection quality. However, faced with a cluttered scene with potential collision, the definition of the grasp-relevant region stays inconsistent, and the relationship between grasps and regional spaces remains incompletely investigated. … Read more

Recommended citation:

Siang Chen, Pengwei Xie, Tang Wei, Dingchang Hu, Yixiang Dai, Guijin Wang. (2024). Region-aware Grasp Framework with Normalized Grasp Space for Efficient 6-DoF Grasping. [pdf]

GAP-RL: Grasps As Points for RL Towards Dynamic Object Grasping

Published in IEEE Robotics and Automation Letters, 2024

Dynamic grasping of moving objects in complex, continuous motion scenarios remains challenging. Reinforcement Learning (RL) has been applied in various robotic manipulation tasks, benefiting from its closed-loop property. … Read more

Recommended citation:

Pengwei Xie, Siang Chen, Qianrun Chen, Wei Tang, Dingchang Hu, Yixiang Dai, Rui Chen, Guijin Wang. (2024). GAP-RL: Grasps As Points for RL Towards Dynamic Object Grasping. [pdf]

talks

teaching

Image Processing

Teaching Assistant in undergraduate course, Department of Electronic Engineering, Tsinghua University, 2022

Image engineering is a new interdisciplinary field that systematically studies various image theories, technologies, and applications. Its content can be divided into three layers: image processing, image analysis, and image understanding. This course is the first one in the image course, mainly introducing low-level content, with a certain depth and breadth. Through the study of this course, readers can solve some practical application problems of image technology on the one hand, and on the other hand, it will also lay a foundation for further learning of mid to high-level technologies in image engineering. Read more

Pattern Recognition

Teaching Assistant in postgraduate course, Department of Electronic Engineering, Tsinghua University, 2023

A course designed for non pattern recognition and intelligent systems majors. Taking rich application examples as the main thread, teach the basic concepts, methods, applications, limitations, and latest developments in modeling, training, classification, and clustering of statistics and structural pattern recognition. Read more

Management Assistant

Management Assistant Teaching Labs of Electronic Engineering Department, Department of Electronic Engineering, Tsinghua University, 2024

Assistant Manager of Teaching Labs of Electronic Engineering Department, Tsinghua University, responsible for the duty of the computer laboratory and maintenance of the servers. Read more