Zhongchen Shi

Zhongchen Shi

Auto-generated profile, based on publicly available information

Zhongchen Shi

Academy of Military Sciences (AMS), Defense Innovation Institute, Beijing, China

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
6 January 2020 in IEEE Access
Verified
Journal Article
Open access
GWGang Wang
Gang Wang
XSXiaoliang Sun
Xiaoliang Sun
YSYang Shang
Yang Shang
ZWZi Wang
Zi Wang
ZSZhongchen Shi
Zhongchen Shi
QYQifeng Yu
Qifeng Yu
The random sample consensus (RANSAC) based algorithm is widely used in estimating the two-view geometry from image point correspondences. However, it often becomes extremely slow when the data is contaminated by a large percentage of incorrect matches. To address this problem, the paper proposes a new modification of RANSAC called LP-RANSAC that is robust to varying inlier ratios and achieves large computational savings without deterioration in accuracy. LP-RANSAC integrates the locality preserving constraint into the universal RANSAC framework, which prunes most of the unreliable correspondences before the hypothesize-and-verify loop and guides non-uniform sampling to generate and verify promising models earlier. Unlike other guided sampling strategies, the proposed method is simple to implement and does not require any prior information. Extensive experiments performed on the publicly available datasets reveal that LP-RANSAC can achieve more accurate and stable solutions at much lower computational cost (in milliseconds on standard CPU) than state-of-the-art methods, particularly when handling problems with low inlier ratios.

19 February 2024 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
MLMeng Liu
Meng Liu
LXLiang Xie
Liang Xie
WWWei Wang
Wei Wang
ZSZhongchen Shi
Zhongchen Shi
WCWei Chen
Wei Chen
EYErWei Yin
ErWei Yin
Augmented reality (AR), a prominent application within the Internet of Things (IoT) domain, demands high-performance pose estimation. Presently, the visual–inertial navigation system (VINS) is acknowledged as an essential method for providing 6-DoF poses. However, VINS builds the local frame at random during the system initialization stage, making it difficult to establish a connection with the global frame. In addition, VINS is prone to drifting. In this article, we propose an innovative method that tightly couples markerless motion capture (Mocap) with vision and an inertial measurement unit (IMU) to achieve global and drift-free pose estimation for AR glasses. To address the issue of pose initialization and establish a connection between the IMU and Mocap, we introduce a coarse-to-fine initialization strategy, enabling data fusion for Mocap, vision, and the IMU under a unified global frame. Furthermore, we formulate the Mocap factor alongside the visual and inertial factors and integrate them into a factor graph framework to constrain the system states. With a spatiotemporal calibration method, the IMU–Mocap extrinsic parameter and time offset are calibrated online to improve the pose estimation accuracy. Experimental evaluations in real-world experiments demonstrate the capability of our method to accurately estimate drift-free poses in the global frame. Compared to the state-of-the-art VINS-Fusion, ORB-SLAM3, and GVIS, we achieve improvements of 81%, 42%, and 33% in translation accuracy and improvements of 58%, 33%, and 72% in rotation accuracy, respectively. Moreover, we also evaluate our system for the EuRoC data set, further indicating the effectiveness of the proposed work.

26 March 2025 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
YLYan Li
Yan Li
ZSZhongchen Shi
Zhongchen Shi
YHYanqing Hou
Yanqing Hou
LXLiang Xie
Liang Xie
HCHongbo Chen
Hongbo Chen
EYErWei Yin
ErWei Yin
Neural Inertial Localization has shown its advantage in simple smartphone usage scenarios, however dominance learning and the model’s adaptability in intricate scenarios have been left unexplored sufficiently. In this paper, we propose a novel implicit self-augmentation and soft dominance prediction method for pedestrian localization, called S2PL. The core idea revolves around a dual-stage learning strategy: 1) a soft dominance predictor (SDP) is firstly devised to achieve dominance-aware learning and identify dominant features of different sensors at each timestamp t. Afterwards, we extend ResMixer to capture spatial contextual information from IMU measurements, followed by combining dominant features to derive the corresponding robust hybrid representations. 2) Introducing implicit self-augmentation technology (i.e., Mixup) to fine-tune our proposed model for better adaptability. Extensive experiments on three real-world datasets demonstrate that the proposed S2PL obtains substantial performance gain over 9 up-to-date state-of-the-art alternatives and makes more than 15% improvements under several scenarios. The real-world inertial tracking trials conducted at the TAIIC campus serve as further validation of the superior generalization capabilities exhibited by the our proposed method.

26 October 2020 in Applied Optics
Verified
Journal Article
Closed access
ZSZhongchen Shi
Zhongchen Shi
KXKai Xu
Kai Xu
ZLZhang Li
Zhang Li
BGBanglei Guan
Banglei Guan
GWGang Wang
Gang Wang
YSYang Shang
Yang Shang
Pose estimation is important for many robotic applications including bin picking and robotic assembly and collaboration. However, robust and accurate estimation of the poses of industrial objects is a challenging task owing to the various object shapes and complex working environments. This paper presents a method of estimating the poses of narrow and elongated industrial objects with a low-cost RGB-D (depth and color) camera to guide the process of robotic assembly. The proposed method comprises three main steps: reconstruction involved in preprocessing, pose initialization with geometric features, and tracking aided by contour cues. Pose tracking is coupled with real-time dense reconstruction, which can synthesize a smooth depth image as a substitute for the raw depth image. Because industrial objects (e.g., fork and adapter) feature mostly planar structures, primitive geometric features, such as three-dimensional planes, are extracted from the point cloud and utilized to induce a promising initial pose. For robust tracking of the adapter consisting of narrow and elongated planes, the dense surface correspondences are combined with sparse contour correspondences in the refinement scheme. This combination allows for a satisfactory tolerance to the initial guess in the pose tracking phase. The experimental results demonstrate the feasibility of the proposed method.

28 October 2025 in IEEE Sensors Journal
Verified
Journal Article
Closed access
ZLZongxuan Li
Zongxuan Li
CCChao Chen
Chao Chen
ZSZhongchen Shi
Zhongchen Shi
WCWei Chen
Wei Chen
LXLiang Xie
Liang Xie
EYErWei Yin
ErWei Yin
Visual–inertial odometry (VIO) is crucial in robotics and augmented reality (AR) applications. Filterbased VIO methods are favored for their lightweight, low memory consumption, and real-time performance properties. However, filter-based methods rely on local visual features for drift correction, and their performance is severely hampered by erroneous visual data during state estimation. To mitigate the impact of outliers on state updates, we propose an adaptive weighting strategy that detects potential outliers and dynamically adjusts their corresponding weights. The simple dynamic adaptation maintains real-time computational efficiency and demonstrates superior robustness to static weighting fashion in visually degraded environments. In addition, temporal constraints are applied to all observations within the sliding window to further improve prediction accuracy. Finally, the proposed method is integrated into a multistate constrained Kalman filter (MSCKF). Our method is implemented within the open-source visual-inertial navigation system (OpenVINS) framework and extensively tested using two well-known public datasets, EuRoC micro aerial vehicle dataset (EuRoC) and TUM visual-inertial dataset (TUM-VI), as well as real-world experiments. Experiments show that our method outperforms existing methods in most sequences. Compared with the baseline method, our method achieves a 36% improvement in average precision on the EuRoC dataset and a 35% improvement in average precision on the TUM-VI dataset.

6 December 2025 in Computers & Graphics
Verified
Journal Article
Closed access
XLXiaohui Li
Xiaohui Li
XLXiaolong Liu
Xiaolong Liu
ZSZhongchen Shi
Zhongchen Shi
WCWei Chen
Wei Chen
LXLiang Xie
Liang Xie
EYErWei Yin
ErWei Yin