WCWenqing Cheng

Wenqing Cheng

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Wenqing Cheng has a total of 77 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Dual-Strategy-Enhanced ConBiMamba for Neural Speaker Diarization

    by Zhen Liao, Gaole Dai, Mengqiao Chen, Wenqing Cheng, Wei Xu

    Conformer and Mamba have achieved strong performance in speech modeling but face limitations in speaker diarization. Mamba is efficient but struggles with local details and nonlinear patterns. Conformer's self-attention incurs high memory overhead for long speech sequences and may cause instability in long-range dependency modeling. These limitations are critical for diarization, which requires both precise modeling of local variations and robust speaker consistency over extended spans. To address these challenges, we first apply ConBiMamba for speaker diarization. We follow the Pyannote pipeline and propose the Dual-Strategy-Enhanced ConBiMamba neural speaker diarization system. ConBiMamba integrates the strengths of Conformer and Mamba, where Conformer's convolutional and feed-forward structures are utilized to improve local feature extraction. By replacing Conformer's self-attention with ExtBiMamba, ConBiMamba efficiently handles long audio sequences while alleviating the high memory cost of self-attention. Furthermore, to address the problem of the higher DER around speaker change points, we introduce the Boundary-Enhanced Transition Loss to enhance the detection of speaker change points. We also propose Layer-wise Feature Aggregation to enhance the utilization of multi-layer representations. The system is evaluated on six diarization datasets and achieves state-of-the-art performance on four of them. The source code of our study is available at https://github.com/lz-hust/DSE-CBM.

  • HIP: Hierarchical Point Modeling and Pre-training for Visual Information Extraction

    by Rujiao Long, Pengfei Wang, Zhibo Yang, Wenqing Cheng

  • Generative compositor for few-shot visual information extraction

    by Zhibo Yang, Wei Hua, Sibo Song, Cong Yao, Yingying Zhu, Wenqing Cheng, Xiang Bai


  • Open Access

    Generative Compositor for Few-Shot Visual Information Extraction

    by Zhibo Yang, Wei Hua, Sibo Song, Cong Yao, Yingying Zhu, Wenqing Cheng, Xiang Bai

    Visual Information Extraction (VIE), aiming at extracting structured information from visually rich document images, plays a pivotal role in document processing. Considering various layouts, semantic scopes, and languages, VIE encompasses an extensive range of types, potentially numbering in the thousands. However, many of these types suffer from a lack of training data, which poses significant challenges. In this paper, we propose a novel generative model, named Generative Compositor, to address the challenge of few-shot VIE. The Generative Compositor is a hybrid pointer-generator network that emulates the operations of a compositor by retrieving words from the source text and assembling them based on the provided prompts. Furthermore, three pre-training strategies are employed to enhance the model's perception of spatial context information. Besides, a prompt-aware resampler is specially designed to enable efficient matching by leveraging the entity-semantic prior contained in prompts. The introduction of the prompt-based retrieval mechanism and the pre-training strategies enable the model to acquire more effective spatial and semantic clues with limited training samples. Experiments demonstrate that the proposed method achieves highly competitive results in the full-sample training, while notably outperforms the baseline in the 1-shot, 5-shot, and 10-shot settings.

  • Open Access

    VL-Reader: Vision and Language Reconstructor is an Effective Scene Text Recognizer

    by Humen Zhong, ZhiBo Yang, Zhaohai Li, Peng Wang, Jun Tang, Wenqing Cheng, Cong Yao

    Text recognition is an inherent integration of vision and language, encompassing the visual texture in stroke patterns and the semantic context among the character sequences. Towards advanced text recognition, there are three key challenges: (1) an encoder capable of representing the visual and semantic distributions; (2) a decoder that ensures the alignment between vision and semantics; and (3) consistency in the framework during pre-training, if it exists, and fine-tuning. Inspired by masked autoencoding, a successful pre-training strategy in both vision and language, we propose an innovative scene text recognition approach, named VL-Reader. The novelty of the VL-Reader lies in the pervasive interplay between vision and language throughout the entire process. Concretely, we first introduce a Masked Visual-Linguistic Reconstruction (MVLR) objective, which aims at simultaneously modeling visual and linguistic information. Then, we design a Masked Visual-Linguistic Decoder (MVLD) to further leverage masked vision-language context and achieve bi-modal feature interaction. The architecture of VL-Reader maintains consistency from pre-training to fine-tuning. In the pre-training stage, VL-Reader reconstructs both masked visual and text tokens, while in the fine-tuning stage, the network degrades to reconstruct all characters from an image without any masked regions. VL-reader achieves an average accuracy of 97.1% on six typical datasets, surpassing the SOTA by 1.1%. The improvement was even more significant on challenging datasets. The results demonstrate that vision and language reconstructor can serve as an effective scene text recognizer.


  • Open Access

    A New Visualization Framework for Multimodal Programming Attention in Debugging Test

    by Jiayi Zeng, Wei Liu, Ying Zhou, Xingyu Li, Wenqing Cheng, Qi Ruan

    The programmer’s attention have been utilized for the behavioral research in programming education. Existing approaches mainly focused on the programmer’s visual attention measured by gaze trackers. However, for the scenarios of programming and debugging, the measurement on just eye movements is not sufficient for understanding programming attention. The measurements on programmer’s other modals of behaviors, including mouse clicks and keyboard inputs, are also required. In this paper, we focus on the programming attention visualization problem in the debugging test scenario. A series of attention visualization approaches are proposed for the two main types of interactions, i.e. reading source code and editing/debugging by keyboard and mouse. We develop methods to visualize student’s attentions from either reading or debugging interactions, and then merge them into multimodal programming attention. The experiment results show that, the proposed multimodal attention graphs can show the diversity of students’ debugging strategies in terms of reading and editing focus in code line and time dimensions. It can help the teacher infer the students’ behavioral intentions and conduct individual diagnose in debugging tests.

  • A Dual-Path Approach for Gaze Following in Fisheye Meeting Scenes

    by Long Rao, Xinghao Huang, Shipeng Cai, Bowen Tian, Wei Xu, Wenqing Cheng


  • Top Cited Papers
    Open Access

    YOLOP: You Only Look Once for Panoptic Driving Perception

    by Dong Wu, Man-Wen Liao, Wei-Tian Zhang, Xing-Gang Wang, Xiang Bai, Wen-Qing Cheng, Wen-Yu Liu

    A panoptic driving perception system is an essential part of autonomous driving. A high-precision and real-time perception system can assist the vehicle in making reasonable decisions while driving. We present a panoptic driving perception network (you only look once for panoptic (YOLOP)) to perform traffic object detection, drivable area segmentation, and lane detection simultaneously. It is composed of one encoder for feature extraction and three decoders to handle the specific tasks. Our model performs extremely well on the challenging BDD100K dataset, achieving state-of-the-art on all three tasks in terms of accuracy and speed. Besides, we verify the effectiveness of our multi-task learning model for joint training via ablative studies. To our best knowledge, this is the first work that can process these three visual perception tasks simultaneously in real-time on an embedded device Jetson TX2(23 FPS), and maintain excellent accuracy. To facilitate further research, the source codes and pre-trained models are released at https://github.com/hustvl/YOLOP.


  • Open Access

    A Simple Optical Aerosol Sensing Method of Sauter Mean Diameter for Particulate Matter Monitoring

    by Liangbo Li, Ang Chen, Tian Deng, Jin Zeng, Feifan Xu, Shu Yan, Shu Wang, Wenqing Cheng, Ming Zhu, Wenbo Xu

    Mass concentration is a commonly used but insufficient metric to evaluate the particulate matter (PM) exposure hazard. Recent studies have declared that small particles have more serious impacts on human health than big particles given the same mass concentration. However, state-of-the-art PM sensors cannot provide explicit information of the particle size for further analysis. In this work, we adopt Sauter mean diameter (SMD) as a key metric to reflect the particle size besides the mass concentration. To measure SMD, an effective optical sensing method and a proof-of-concept prototype sensor are proposed by using dual wavelengths technology. In the proposed method, a non-linear conversion model is developed to improve the SMD measurement accuracy for aerosol samples of different particle size distributions and reflective indices based on multiple scattering channels. In the experiment of Di-Ethyl-Hexyl-Sebacate (DEHS) aerosols, the outputs of our prototype sensor demonstrated a good agreement with existing laboratory reference instruments with maximum SMD measurement error down to 7.04%. Furthermore, the simplicity, feasibility and low-cost features of this new method present great potential for distributed PM monitoring, to support sophisticated human exposure hazard assessment.