ZHZiyang Hong

Ziyang Hong

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Ziyang Hong has a total of 12 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Real-Time 3D Visual Perception by Cross-Dimensional Refined Learning

    by Ziyang Hong, C. Patrick Yue

    We introduce a novel learning method that can effectively perceive both the geometry structure and semantic labels of a 3D scene in real time. Existing real-time 3D scene reconstruction approaches often rely on volumetric schemes to regress a Truncated Signed Distance Function (TSDF) as the 3D representation. However, these volumetric approaches primarily focus on ensuring global coherence in the reconstructed scene, which often results in a lack of local geometric detail. To address this limitation, we propose a solution that leverages the latent geometric knowledge present in 2D image features by explicit depth prediction thereby creating anchored features, which are used to refine the learning of occupancy in the TSDF volume. Furthermore, we discover that this cross-dimensional feature refinement methodology can also be applied to the task of semantic segmentation by utilizing semantic priors. As a result, we propose an end-to-end cross-dimensional refinement neural network (CDRNet) that can extract both the 3D mesh and 3D semantic labeling of a scene in real time. Through experimental evaluation on multiple datasets, we demonstrate that our method achieves state-of-the-art 3D perception capability by boosting over 40% and 18% in 3D semantic segmentation and geometric reconstruction respectively over the prior art. These promising results indicate the significant potential of our approach for various industrial applications. Demo video and code can be found on the project page, https://hafred.github.io/cdrnet/.


  • Interpretable image classification method for industrial bearing based on GWO-AlexNet

    by Jian Gao, Xin Wang, Ziyang Hong

    In response to the problem of bearing fault diagnosis, a deep learning method based on the combination of GWO-AlexNet group optimization algorithm and convolutional neural network, and the use of Grad-CAM for decision interpretability, is proposed. Firstly, the original vibration signal data of the bearing is transformed into images for preprocessing, and the corresponding target dataset is generated; Then train and fine tune the pre trained GWO-AlexNet model on the generated target dataset; Finally, the fine-tuned GWO-AlexNet model will be applied to fault diagnosis training and decision interpretation. The method was validated on the bearing dataset of Case Western Reserve University, and the results showed that it can achieve nearly 100% accuracy in bearing fault diagnosis and can effectively interpret diagnostic decisions, with good application prospects.

  • Open Access

    Efficient-Grad: Efficient Training Deep Convolutional Neural Networks on Edge Devices with Grad ient Optimizations

    by Ziyang Hong, C. Patrick Yue

    With the prospering of mobile devices, the distributed learning approach, enabling model training with decentralized data, has attracted great interest from researchers. However, the lack of training capability for edge devices significantly limits the energy efficiency of distributed learning in real life. This article describes Efficient-Grad, an algorithm-hardware co-design approach for training deep convolutional neural networks, which improves both throughput and energy saving during model training, with negligible validation accuracy loss. The key to Efficient-Grad is its exploitation of two observations. Firstly, the sparsity has potential for not only activation and weight, but gradients and the asymmetry residing in the gradients for the conventional back propagation (BP). Secondly, a dedicated hardware architecture for sparsity utilization and efficient data movement can be optimized to support the Efficient-Grad algorithm in a scalable manner. To the best of our knowledge, Efficient-Grad is the first approach that successfully adopts a feedback-alignment (FA)-based gradient optimization scheme for deep convolutional neural network training, which leads to its superiority in terms of energy efficiency. We present case studies to demonstrate that the Efficient-Grad design outperforms the prior arts by 3.72x in terms of energy efficiency.


  • Collocated and Simultaneous Measurements of RF Current and Voltage on a Trace in a Noncontact Manner

    by Chengyang Luo, Wenxiao Fang, Yang Shen, Haimi Qiu, Weiheng Shao, Ziyang Hong, Lei Wang, Zhiyuan He, Yunfei En

    With the trend of miniature, high-integrated assembly modules are common in modern industry. The measurement of the flowing current and voltage on a trace is basic for diagnosis and monitoring of these modules. The traditional contact detection method is not practical especially in conditions of high power and small space. In this paper, we demonstrate a novel method for collocated and simultaneous measurements of radio frequency current and voltage on a trace in a noncontact and movable manner. The measurements are achieved by using a compact probe with both voltage and current detectors, as well as an algorithm based on the transfer functions quantifying the capacitive and inductive couplings between the detectors and the tested trace. The proposed method is verified and applied for continuous power measurements, collocated and simultaneous measurements of transient voltage and current pulses and the dynamic voltage and current of a switching device.


  • Electromagnetic Pattern Extraction and Grouping for Near-Field Scanning of Integrated Circuits by PCA and K-Means Approaches

    by Ziyang Hong, Wenxiao Fang, Yunfei En, Chengyang Luo, Weiheng Shao, Lei Wang, Zhiyuan He, E. Shao

    Electromagnetic patterns in near field scanning usually give important information about the source inside an device under test. In this paper, we develop a postprocessing technique, including standard-deviation-based method, principal component analysis, and K-means method, so that the electromagnetic patterns in near field scanning can be extracted fast and grouped automatically and efficiently. The postprocessing techniques are applied to experimental near-field scanning of ICs, such as a microcontroller unit (MCU) and a field programmable gate array (FPGA). The grouping on the electromagnetic patterns of the MCU is in good agreement with the published grouping result, and the grouping on the electromagnetic patterns of the FPGA can find out the additional cluster responsible for additional circuit.