HWHuiqian Wang

Huiqian Wang

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Huiqian Wang has a total of 53 co-authors. Here are the top co-authors by number of shared publications:

  • Yu Pang

    School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing, China

    14 shared publications
  • Jinzhao Lin

    School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing, China

    5 shared publications
  • En Mou

    School of Communication and Information Engineering, Chongqing University of Posts and Telecommunication, Chongqing 400065, China

    3 shared publications
  • Enling Cao

    School of Software Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China

    3 shared publications
  • Jiasai Luo

    School of Optoelectronic Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China

    3 shared publications
  • Tong Bai

    Chongqing Key Laboratory of Photoelectronic Information Sensing and Transmitting Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China

    3 shared publications

Latest Publications

  • A Lightweight UAV Object Detector Based on Optimized YOLOv8 Fused With an Auxiliary Learning Branch for AIoT

    by Yuanyuan Chen, Huiqian Wang, Zhangyong Li, En Mou, Tiecheng Song, Shuai Xia, Yu Pang

    In Artificial Intelligence of Things (AIoT)-enabled edge sensing systems, uncrewed aerial vehicles (UAVs) have emerged as the dominant low-altitude monitoring platform owing to their wide-area coverage. However, in large-scale surveillance scenarios, sensor limitations and the inherent difficulty in preserving discriminative features for tiny targets hinder robust small-object detection. Therefore, we propose the multiscale feature reconstruction-densely efficient flow-you only look once (MSFR-DEF-YOLO) framework, which combines DEF-YOLO with MSFR auxiliary branches. We first enhance fine-grained feature extraction by optimizing the network structure and integrating an efficient downsampling module (EDM). Next, a dynamic dense feature pyramid (DDFP) is designed to incorporate dense cross-layer connections and bidirectional feature flows, enhancing multiscale context consistency. In addition, the introduction of the MSFR auxiliary branch integrates shallow and deep features from the backbone network, enabling feature reconstruction (FR) and information compensation, thereby further improving small-object detection capabilities. To verify the improvements, we tested a range of network configurations on different YOLO versions. Our model outperforms other versions in the YOLO series. Moreover, on four public datasets, MSFR-DEF-YOLO achieves mAP50 improvements of 8.8%, 6.9%, 1.8%, and 6% compared to YOLOv8n, all while using only 20% of the parameters of YOLOv8n. To further reduce the computational overhead, this article proposes a lightweight MSFR-L-YOLO model that utilizes a unidirectional path neck design. Compared to MSFR-DEF-YOLO, this model reduces the number of parameters by 15% while maintaining high efficiency, thereby reducing computational requirements and achieving comparable accuracy. This provides an efficient solution for AIoT-enabled airborne sensing systems.


  • Low-Light Endoscopic Image Enhancement for Healthcare Electronics Using Efficient Multiscale Selective Fusion

    by En Mou, Huiqian Wang, Xiaodong Chen, Zhangyong Li, Lisha Zhong, Shuai Xia, Guanjie Zeng, Yu Pang

    As essential healthcare electronics, endoscopy systems suffer from non-uniform illumination due to factors such as their built-in lighting configuration, potentially obscuring critical tissue landmarks and subtle pathologies, thereby reducing diagnostic confidence, especially in deep cavities. To address these issues, endoscopic image enhancement methods are required to preserve structural details and improve tissue contrast while maintaining low computational cost. Convolution-based methods enhance global brightness at the cost of attenuating high-frequency textures due to spectral smoothing effects. To overcome this limitation, we propose a multiscale channel selective fusion network to enhance endoscopic images by discriminatively fusing synergistic features via gated multiscale attention, simultaneously optimizing local discriminability and global consistency. Specifically, a selective fusion module is designed to effectively extract and fuse channel information symmetrically at each scale, enhancing brightness and preserving fine textures. Experiments on the Endo4IE and LEIED endoscopic datasets demonstrate that the proposed method outperforms existing nine state-of-the-art techniques under low-light conditions. It offers an effective method for low-light endoscopic image enhancement in healthcare electronics.


  • Open Access

    Retinex theory-based nonlinear luminance enhancement and denoising for low-light endoscopic images

    by En Mou, Huiqian Wang, Xiaodong Chen, Zhangyong Li, Enling Cao, Yuanyuan Chen, Zhiwei Huang, Yu Pang

    The quality of low-light endoscopic images involves applications in medical disciplines such as physiology and anatomy for the identification and judgement of tissue structures. Due to the use of point light sources and the constraints of narrow physiological structures, medical endoscopic images display uneven brightness, low contrast, and a lack of texture information, presenting diagnostic challenges for physicians. In this paper, a nonlinear brightness enhancement and denoising network based on Retinex theory is designed to improve the brightness and details of low-light endoscopic images. The nonlinear luminance enhancement module uses higher-order curvilinear functions to improve overall brightness; the dual-attention denoising module captures detailed features of anatomical structures; and the color loss function mitigates color distortion. Experimental results on the Endo4IE dataset demonstrate that the proposed method outperforms existing state-of-the-art methods in terms of Peak Signal-to-Noise Ratio (PSNR), Structural Similarity (SSIM), and Learned Perceptual Image Patch Similarity (LPIPS). The PSNR is 27.2202, SSIM is 0.8342, and the LPIPS is 0.1492. It provides a method to enhance image quality in clinical diagnosis and treatment. It offers an efficient method to enhance images captured by endoscopes and offers valuable insights into intricate human physiological structures, which can effectively assist clinical diagnosis and treatment.


  • Retraction
    Open Access

    RETRACTED: Chen et al. An Infrared Small Target Detection Method Based on a Weighted Human Visual Comparison Mechanism for Safety Monitoring. Remote Sens. 2023, 15, 2922

    by Yuanyuan Chen, Huiqian Wang, Yu Pang, Jinhui Han, En Mou, Enling Cao

    The journal retracts the article titled “An Infrared Small Target Detection Method Based on a Weighted Human Visual Comparison Mechanism for Safety Monitoring”


  • Open Access

    Global and Local Enhancement of Low-light Endoscopic Images

    by En Mou, Huiqian Wang, Meifang Yang, Enling Cao, Yuanyuan Chen, Chunlan Ran, Yu Pang

    Endoscopic medical images can suffer from uneven illumination, low contrast, and lack of texture information due to the use of point directional light sources and the presence of narrow tissue structures, posing diagnostic difficulties for physicians. In this paper, a deep learning-based su-pervised illumination enhancement network is designed for low-light endoscopic images, aiming to improve both global illumination and local details. Initially, a global illumination enhancement module is formulated utilizing a higher-order curve function to improve global illumination. Sec-ondly, a local feature extraction module incorporating dual attention is designed to capture local detailed features. Considering the significance of color fidelity in biomedical scenarios, the designed loss function prioritizes introducing color loss to alleviate image color distortion. Compared with seven state-of-the-art enhancement algorithms on Endo4IE endoscopic datasets, experimental re-sults show that the proposed method can better enhance low-light endoscopic images and avoid image color distortion. It provides an efficient method to enhance images captured by endoscopes which can effectively assist clinical diagnosis and treatment.


  • Retracted
    Open Access

    RETRACTED: An Infrared Small Target Detection Method Based on a Weighted Human Visual Comparison Mechanism for Safety Monitoring

    by Yuanyuan Chen, Huiqian Wang, Yu Pang, Jinhui Han, En Mou, Enling Cao

    Infrared small target detection is a crucial technology in both military and civilian applications, including surveillance, security, defense, and combat. However, accurate infrared detection of small targets in real-time is challenging due to their small size and similarity in gray level and texture with the surrounding environment, as well as interference from the infrared imaging systems in unmanned aerial vehicles (UAVs). This article proposes a weighted local contrast method based on the contrast mechanism of the human visual system. Initially, a combined contrast ratio is defined that stems from the pixel-level divergence between the target and its neighboring pixels. Then, an improved regional intensity level is used to establish a weight function with the concept of ratio difference combination, which can effectively suppress complex backgrounds and random noise. In the final step, the contrast and weight functions are combined to create the final weighted local contrast method (WRDLCM). This method does not require any preconditioning and can enhance the target while suppressing background interference. Additionally, it is capable of detecting small targets even when their scale changes. In the experimental section, our algorithm was compared with some popular methods, and the experimental findings indicated that our method showed strong detection capability based on the commonly used performance indicators of the ROC curve, SCRG, and BSF, especially in low signal-to-noise ratio situations. In addition, unlike deep learning, this method is appropriate for small sample sizes and is easy to implement on FPGA hardware.


  • Open Access

    Polarization Orientation Method Based on Remote Sensing Image in Cloudy Weather

    by Jiasai Luo, Sen Zhou, Yiming Li, Yu Pang, Zhengwen Wang, Yi Lu, Huiqian Wang, Tong Bai

    Autonomous navigation technology is a core technology for intelligent operation, allowing the vehicles to perform tasks without relying on external information, which effectively improves the concealability and reliability. In this paper, based on the previous research on the bionic compound eye, a multi-channel camera array with different polarization degrees was used to construct the atmospheric polarization state measurement platform. A polarization trough threshold segmentation algorithm was applied to study the distribution characteristics and characterization methods of polarization states in atmospheric remote sensing images. In the extracted polarization feature map, the tilting suggestion box was obtained based on the multi-direction window extraction network (similarity-based region proposal networks, SRPN) and the rotation of the suggestion box (Rotation Region of interests, RRoIs). Fast Region Convolutional Neural Networks (RCNN) was used to screen the suggestion boxes, and the Non-maximum suppression (NMS) method was used to select the angle, corresponding to the label of the suggestion box with the highest score, as the solar meridian azimuth in the vehicle coordinate system. The azimuth angle of the solar meridian in the atmospheric coordinate system can be calculated by the astronomical formula. Finally, the final heading angle can be obtained according to the conversion relationship between the coordinate systems. By fitting the measured data based on the least Square method, the slope K value is −1.062, RMSE (Root Mean Square Error) is 6.984, and the determination coefficient R-Square is 0.9968. Experimental results prove the effectiveness of the proposed algorithm, and this study can construct an autonomous navigation algorithm with high concealment and precision, providing a new research idea for the research of autonomous navigation technology.


  • Open Access

    Deep-Learning-Based Complex Scene Text Detection Algorithm for Architectural Images

    by Weiwei Sun, Huiqian Wang, Yi Lu, Jiasai Luo, Ting Liu, Jinzhao Lin, Yu Pang, Guo Zhang

    With the advent of smart cities, the text information in an image can be accurately located and recognized, and then applied to the fields of instant translation, image retrieval, card surface information recognition, and license plate recognition. Thus, people’s lives and work will become more convenient and comfortable. Owing to the varied orientations, angles, and shapes of text, identifying textual features from images is challenging. Therefore, we propose an improved EAST detector algorithm for detecting and recognizing slanted text in images. The proposed algorithm uses reinforcement learning to train a recurrent neural network controller. The optimal fully convolutional neural network structure is selected, and multi-scale features of text are extracted. After importing this information into the output module, the Generalized Intersection over Union algorithm is used to enhance the regression effect of the text bounding box. Next, the loss function is adjusted to ensure a balance between positive and negative sample classes before outputting the improved text detection results. Experimental results indicate that the proposed algorithm can address the problem of category homogenization and improve the low recall rate in target detection. When compared with other image detection algorithms, the proposed algorithm can better identify slanted text in natural scene images. Finally, its ability to recognize text in complex environments is also excellent.


  • Research on text detection on building surfaces in smart cities based on deep learning

    by Yuanpeng Long, Weiwei Sun, Yu Pang, Huiqian Wang, Guo Zhang


  • A desmoking algorithm for endoscopic images based on improved U‐Net model

    by Jinzhao Lin, Meiqiu Jiang, Yu Pang, Huiqian Wang, Zhu Chen, Chongyuan Yan, Qinghui Liu, Yuanfa Wang

    In laparoscopic surgery, the smoke generated by operations including electrocautery and laser ablation seriously degrades the quality of endoscopic images. It not only reduces the visibility of the surgery, leading to increased risk of surgery, but also affects the performance of image processing in computer‐assisted surgery such as segmentation, 3D reconstruction and tracking. Therefore, a desmoking algorithm is required to eliminate smoke in endoscopic images. In this article, we study a U‐Net model that can eliminate smoke of the laparoscopic image in real‐time and preserve the natural appearance of the organ surface. Our method is based on the improved U‐Net in which convolutional block attention module is used as an embedded guide mask of the decoder part. The laparoscopic image dataset is provided by the Hamlyn Center, and the Blender software is used to simulate various situations of smoke added to the laparoscopic image for training and testing. For the proposed method, the peak signal‐to‐noise ratio value is up to 29.27 and the structural similarity index is up to 0.945 over the test images. The experimental results demonstrate that the proposed method achieves a better performance than other six existing methods, which is applicable for real‐time endoscopic image processing.