CZCheng Zhou

Cheng Zhou

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Cheng Zhou has a total of 11 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • A survey on large model-driven embodied grasping technologies

    by Xu Lu, Haijun Liu, Hanyuan Huang, Cheng Zhou, Fujian Yu, Xinyu Wu, Jun Liu

    Robotic grasping is a crucial step in enabling practical robotic applications, and many scholars in the robotics field have conducted long-term and in-depth studies in this area. In recent years, embodied intelligent systems driven by large models have demonstrated remarkable potential in robotic grasping tasks. By integrating multimodal perception, semantic understanding, reasoning ability, and complex decision-making capabilities, they have promoted the evolution of grasping technology from simple action execution toward generalization, adaptability, and robustness across diverse, dynamic, and unstructured environments. These advances have further highlighted the importance of combining data-driven intelligence with physical priors to enhance both efficiency and safety. This paper reviews and summarizes key technologies related to large model-driven embodied intelligent grasping, including intelligent perception and scene understanding, grasping task planning and action execution, process optimization, advances in simulation platforms and datasets, as well as current challenges, open questions, and future prospects. The aim is to provide a comprehensive and systematic overview of the field, offering deeper insights, establishing connections across subdomains, and facilitating sustained progress for researchers and practitioners worldwide.

  • GASegNet: Global Self-Attention Mechanism Meets Structural Feature Fusion for Point Cloud Semantic Segmentation

    by Xu Lu, Haijun Liu, Guang'AN Luo, Zhike Chen, Cheng Zhou, Xinyu Wu, Jun Liu

    With the rapid development of autonomous driving technology, semantic segmentation, as one of the key technologies contributing to the environment perception of autonomous driving systems, still suffers from a lack of connections between local features, as well as high computational consumption and an inability to meet real-time requirements. To address the above problems, this paper proposes a lightweight and efficient point cloud semantic segmentation network based on spherical projection with encoder-decoder structure. The encoder consists of a global self-attention mechanism that captures global information, as well as multi-scale convolution. This module achieves the unification of local feature extraction and global characteristic information for high-dimensional semantic information. In order to alleviate the high computational cost, a feature fusion module is introduced to enhance the compactness of the range image structure obtained from point cloud projection. The decoder utilizes bilinear interpolation to upsample multi-resolution feature maps, and introduces multiple auxiliary segmentation heads to further enhance the network's accuracy. Experiments conducted on the SemanticKITTI and SemanticPOSS datasets reveal that, in comparison to the CENet architecture, the proposed approach attains enhancements in mIoU of 4.3% and 2.6% on the respective datasets, thereby substantiating its efficacy. The code is available at GitHub: https://github.com/haifeng925/GASegNet.

  • RSD-SLAM: A Robust Saliency-Driven Visual SLAM System in Indoor Environments

    by Xu Lu, Cheng Zhou, Kejie Zhong, Hanyuan Huang, Zhike Chen, Guang'AN Luo, Jun Liu, Xinyu Wu

    The region of interest (ROI) with abundant and structured textures provides robust features in an indoor environment, which can effectively facilitate accurate simultaneous localization and mapping (SLAM). However, most existing visual SLAM systems generally treat ROI and non-ROI uniformly, resulting in ineffective employment of ROI. To meet this gap, we propose a robust saliency-driven visual SLAM system for indoor environments, coined RSD-SLAM. It can increase the focus on valuable ROI with the saliency maps obtained from a novel saliency prediction (SP) model. Specifically, we first design a saliency map construction method for visual SLAM enabling the SP model to accurately describe ROI, which generates the first indoor SP dataset integrating geometric, semantic, depth, and low-level visual information. Second, we develop a global stability constraint module for the SP model to enable the capability of keeping temporal consistency and illumination invariance. Third, we design a saliency map based hybrid saliency-driven mechanism to increase the focus of the system on ROI. At the front-end of the system, an adaptive feature-point extraction algorithm extracts more robust feature-points from ROI, and a saliency entropy based keyframe selection algorithm selects keyframes with the saliency value distribution of feature points. At the back-end, a dynamic weighted bundle adjustment (BA) optimization algorithm heavily weights the map points of ROI. Last, the particular focus on ROI results in a robust and accurate location. Extensive experiments, conducted on the EuRoC and TUM RGB-D datasets as well as in simulation environments, demonstrate that the proposed RSD-SLAM significantly outperforms the state-of-the-art in robustness and accuracy.


  • Open Access

    AFERSM-Net: Joint Network for Gesture Recognition and Location Classification

    by Xu Lu, Zexiao Cai, Xiongwei Huang, Cheng Zhou, Jun Liu

    With the widespread deployment of wireless communication systems and smart devices, gesture recognition and indoor location classification technologies based on WiFi wireless devices are increasingly used. Its technical principle is to identify human activities and locations by extracting gesture and location features from WiFi channel state information (CSI). However, the signal is susceptible to interference from the environment during CSI data acquisition to produce multipath effect noise, and the amplitude change with the change of location often affects the extraction and recognition of gesture features. To address these problems, Auxiliary Feature Extraction based Residual Shrinkage Multi-tasking Network (AFERSM-Net) is proposed for gesture recognition and position classification of one-dimensional multivariate time series. AFERSM-Net is a hybrid architecture that combines CNN for spatial feature extraction and LSTM networks for capturing temporal dependencies. Firstly, a reasonable threshold is set adaptively by the shrinkage module to dynamically identify and eliminate the transformed environmental noise. Secondly, the feature extraction module is used to focus on and extract location-independent gesture features to reduce the influence of location-independent features. Finally, the gesture features extracted by the feature extraction module are fused with the shared features of the residual shrinkage multi-tasking network as an aid. Its module fusion is mainly used to improve the accuracy of gesture recognition and solve the problem of insufficient model generalization ability. We evaluated this network on a dual-labeled gesture and location dataset, an the gesture recognition accuracy was 97.84% and the location classification accuracy was 98.92%, which outperformed other advanced network frameworks.