Query：
学者姓名：吴林煌
Refining：
Year
Type
Indexed by
Source
Complex
Former Name
Co-
Language
Clean All
Abstract ：
针对当前基于深度学习的行人检测算法检测精度低、模型复杂、对设备要求较高的问题,提出一种基于YOLOv8 的轻量级行人检测模型YOLOv8-PGL.首先,将卷积神经网络(Convolutional Neural Networks,CNN)和Transformer的思想结合,设计出C2f_PTB模块,增强特征提取能力,降低计算量和参数量;其次,引入改进的BiFPN网络替换原模型中的特征提取网络,提高对不同尺度特征的融合效率;再次,采用一种轻量级非对称检测头LADH,以微小性能损失大幅度减少模型的计算量和参数量;最后,使用PIoU(Powerful-IoU)替换CIoU(Complete-IoU)作为损失函数,以更准确地优化模型的预测结果,进一步提高模型的检测精度.实验结果表明,所提出的模型在多个指标上有明显的提升.相较于基准模型,YOLOv8-PGL的mAP50%提升1.9个百分点,参数量降低50%,计算量降低33%,模型大小降低48%.
Keyword ：
YOLOv8 YOLOv8 深度学习 深度学习 行人检测 行人检测 轻量化 轻量化
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 苏喆 , 徐成也 , 吴林煌 . 轻量级行人检测算法研究 [J]. | 电视技术 , 2025 , 49 (2) : 30-36 . |
| MLA | 苏喆 等. "轻量级行人检测算法研究" . | 电视技术 49 . 2 (2025) : 30-36 . |
| APA | 苏喆 , 徐成也 , 吴林煌 . 轻量级行人检测算法研究 . | 电视技术 , 2025 , 49 (2) , 30-36 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
JPEG-XS视频编解码标准具有高质量、低复杂度、低延时等特点.针对JPEG-XS图像编解码压缩标准,对其解码算法进行了简要介绍,提出了一种面向硬件实现的高性能JPEG-XS解码器架构.所设计的解码器硬件架构采用流水线处理,能够在保持高数据吞吐量的同时减少由组合逻辑带来的路径延迟,提高了工作频率,每个时钟周期可解码4个重构像素值.实验结果表明,在Xilinx Zynq FPGA的实验平台上,所设计的高性能JPEG-XS解码器硬件架构仅占用约15×103个查找表和11×103个寄存器资源,最高主频达254 MHz,最高可支持4K、100帧/s的实时视频解码.
Keyword ：
FPGA FPGA JPEG-XS JPEG-XS 硬件架构 硬件架构 视觉无损 视觉无损 解码器 解码器
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 郑畅 , 吴林煌 , 李雅欣 et al. 基于FPGA的JPEG-XS高性能解码器硬件架构设计 [J]. | 电子与封装 , 2025 , 25 (2) : 49-54 . |
| MLA | 郑畅 et al. "基于FPGA的JPEG-XS高性能解码器硬件架构设计" . | 电子与封装 25 . 2 (2025) : 49-54 . |
| APA | 郑畅 , 吴林煌 , 李雅欣 , 刘伟 . 基于FPGA的JPEG-XS高性能解码器硬件架构设计 . | 电子与封装 , 2025 , 25 (2) , 49-54 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
为将JPEG-XS这一主流的浅压缩算法与现场可编程门阵列(FPGA)相结合,设计了一种适用于高分辨率、高帧率应用场景的视频编码器,提出了一种完整的JPEG-XS编码器硬件方案.对整个编码器进行流水线编码设计,实现模块间时间上的复用,对于模块内部,提出了4行并行计算的5/3小波变换架构,对于耗时最长的熵编码模块提出了并行编码各子包的硬件方案.实验结果表明,在Xilinx UltraScale+ZCU102的FPGA平台,该硬件架构仅占用38.9×103个查找表资源和23.8×103个寄存器资源,最大主频可达182.24 MHz,可支持4K@60帧/s的实时编码.
Keyword ：
JPEG-XS JPEG-XS 并行度 并行度 现场可编程门阵列 现场可编程门阵列 硬件架构 硬件架构
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 李雅欣 , 吴林煌 , 刘伟 et al. 一种JPEG-XS编码器的硬件架构优化设计 [J]. | 电子与封装 , 2025 , 25 (2) : 55-61 . |
| MLA | 李雅欣 et al. "一种JPEG-XS编码器的硬件架构优化设计" . | 电子与封装 25 . 2 (2025) : 55-61 . |
| APA | 李雅欣 , 吴林煌 , 刘伟 , 郑畅 . 一种JPEG-XS编码器的硬件架构优化设计 . | 电子与封装 , 2025 , 25 (2) , 55-61 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
为了面向低延时的浅压缩场景提供更加适配的编码方案,并降低硬件实现成本,提出一种基于数字音视频编解码技术标准(Audio Video coding Standard,AVS)浅压缩算法的帧内预测模式优化以及快速率失真优化算法.该算法通过减少原有算法帧内预测所需的预测循环次数,以及打破各块之间的数据依赖关系等措施,克服了原始方案不适合硬件流水并行处理的限制,提高了编码的效率和稳定性,从而既保障了算法的视频质量,又使新的硬件实现方案更符合实际应用需求.实验结果表明,该算法优化方案能够有效改善实际面向低延时浅压缩场景下的编码效果.
Keyword ：
帧内预测 帧内预测 快速算法 快速算法 数字音视频编解码技术标准(AVS) 数字音视频编解码技术标准(AVS) 浅压缩 浅压缩
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 蔡江震 , 吴林煌 , 黄贺焜 . 面向硬件的AVS帧内预测算法改进 [J]. | 电视技术 , 2024 , 48 (5) : 41-47 . |
| MLA | 蔡江震 et al. "面向硬件的AVS帧内预测算法改进" . | 电视技术 48 . 5 (2024) : 41-47 . |
| APA | 蔡江震 , 吴林煌 , 黄贺焜 . 面向硬件的AVS帧内预测算法改进 . | 电视技术 , 2024 , 48 (5) , 41-47 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
In this paper, we valign="-60pt propose a joint differential pulse position modulation and differential chaos shift key modulation (DPPM-DCSK), where one bit modulates the reference PPM signal by chaotic signal, and the other m(c) bits are mapped by the position index of this chaotic pulse in the information signal. In particular, in the DPPM-DCSK, the information-carrying PPM signal for the current symbol also serves as the reference for the next symbol. It can avoid the energy and rate wastes in the conventional PPM-DCSK. Moreover, we redesign the differential Walsh codes (WC) for DPPM-DCSK to enable multi-user communications. The numerical bit-error-rate (BER) performance of the proposed DPPM-DCSK-WC over multipath Rayleigh fading channels are derived and then verified by simulations. The results demonstrate that the DPPM-DCSK-WC can achieve performance gains of more than 2 dB over the conventional PPM-DCSK and DDCSK-WC for multi-user scenarios, and this gain can be up to 5 dB in the high-delay channels at a BER of 10(-4). In addition, the superior performance of the proposed scheme is proved over ultra-wideband (UWB) wireless channels, which indicates its great potential for chaotic-based wireless applications.
Keyword ：
Chaotic communication Chaotic communication differential chaos shift keying (DCSK) differential chaos shift keying (DCSK) multi-access multi-access pulse position modulation(PPM) pulse position modulation(PPM) walsh code (WC) walsh code (WC)
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Wu, Linhuang , Yin, Xiangxiang , Chen, Haoyu et al. Differential Pulse-Position Modulation for Multi-User Chaotic Communication [J]. | IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY , 2024 , 73 (8) : 11303-11317 . |
| MLA | Wu, Linhuang et al. "Differential Pulse-Position Modulation for Multi-User Chaotic Communication" . | IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 73 . 8 (2024) : 11303-11317 . |
| APA | Wu, Linhuang , Yin, Xiangxiang , Chen, Haoyu , Chen, Pingping , Fang, Yi . Differential Pulse-Position Modulation for Multi-User Chaotic Communication . | IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY , 2024 , 73 (8) , 11303-11317 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Detection-based tracking is a prominent multi-object tracking (MOT) paradigm that utilizes powerful detectors for high tracking performance. However, this method primarily relies on object prediction, leading to error accumulation when objects are occluded or lost. To address this issue, we propose a novel spatial clustering-guided MOT method (SCGTrack). First, we introduce a fast spatial clustering (FSC) algorithm to efficiently group the detections by their spatial information to form clusters. Second, we develop a spatial clustering matching module to reconnect lost trajectories and re-match trajectories within the clusters. By doing so, SCGTrack can continuously track the lost objects as long as the cluster exists, which effectively mitigates long-term occlusion problems. Extensive experiment results on the MOT17 and DanceTrack dataset demonstrate that our approach achieves state-of-the-art performance in terms of tracking accuracy, with an exceptional matching speed of over 800 FPS on a single CPU. Code is available at https://github.com/Yamahhh/SCGTrack.
Keyword ：
Multi-object tracking (MOT) Multi-object tracking (MOT) Power capacitors Power capacitors Prediction algorithms Prediction algorithms Radar tracking Radar tracking real-time tracking real-time tracking Sensors Sensors Task analysis Task analysis Tracking Tracking tracklet association tracklet association Trajectory Trajectory unsupervised clustering unsupervised clustering
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Chen, Yulin , Li, Zongtan , Wu, Linhuang et al. An Efficient Multi-Object Tracking Guided by Spatial Clustering on Vision Sensors [J]. | IEEE SENSORS JOURNAL , 2024 , 24 (12) : 19344-19351 . |
| MLA | Chen, Yulin et al. "An Efficient Multi-Object Tracking Guided by Spatial Clustering on Vision Sensors" . | IEEE SENSORS JOURNAL 24 . 12 (2024) : 19344-19351 . |
| APA | Chen, Yulin , Li, Zongtan , Wu, Linhuang , Chen, Pingping . An Efficient Multi-Object Tracking Guided by Spatial Clustering on Vision Sensors . | IEEE SENSORS JOURNAL , 2024 , 24 (12) , 19344-19351 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
The majority of image denoising algorithms assume that the noise is evenly distributed white Gaussian noise, however, the image noise which collected in real scenes is more complex. In this paper, we propose a real image denoising method based on pixel-level noise estimation. The method is improved on the basis of the block-matching and 3D filtering (BM3D) image denoising algorithm. The noise estimation algorithm further introduces pixel-level non-local self similarity (NSS) on the basis of patch-level NSS prior. After detecting the flatness of the image block, the relevant parameters are adaptively adjusted, finally the noise estimation algorithm and the image denoising algorithm are combined block by block. Experiments show that this noise estimation method greatly reduce the required processing time while ensuring the accuracy of noise estimation. The image denoising effect has certain superiority compared with other classic traditional denoising algorithms. © 2022 ACM.
Keyword ：
Gaussian noise (electronic) Gaussian noise (electronic) Image denoising Image denoising Image enhancement Image enhancement Motion compensation Motion compensation Pixels Pixels
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Hong, Hui , He, Mingsen , Wang, Kaixin et al. An Image Denoising Method for Real Scene Based on Pixel-Level Noise Estimation [C] . 2022 : 306-311 . |
| MLA | Hong, Hui et al. "An Image Denoising Method for Real Scene Based on Pixel-Level Noise Estimation" . (2022) : 306-311 . |
| APA | Hong, Hui , He, Mingsen , Wang, Kaixin , Wu, Linhuang . An Image Denoising Method for Real Scene Based on Pixel-Level Noise Estimation . (2022) : 306-311 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
License plate recognition technology plays an important role in traffic management. It is widely used in parking, high-speed and road traffic management. The existing license plate recognition system is easy to be disturbed by the external environment, and the detection performance is poor in the night scene. This paper investigates the problem of license plate recognition in night vision scenarios and license plate tilt. This paper proposes a license plate recognition method with night vision enhancement, which can detect and recognize license plates under extremely poor lighting conditions. The method first uses a night vision enhancement module, Recursive Encoder-Decoder Network (RED-Net), and a set of non-reference loss functions designed for properties of the image. And then the License Plate Location and Recognition (LPLR) system is used to get the license plate number. The algorithm is tested on 75k images of the simulated CCPD night dataset. The testing result that the accuracy of the license plate recognition algorithm after night vision enhancement can reach 72.29% increase 65.5% than without night vision enhancement. © 2022 IEEE.
Keyword ：
Deep learning Deep learning Highway planning Highway planning Highway traffic control Highway traffic control Image enhancement Image enhancement Intelligent systems Intelligent systems License plates (automobile) License plates (automobile) Optical character recognition Optical character recognition Vision Vision
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Wang, Kaixin , Wu, Linhuang , Zhang, Shihao et al. Night Vision Enhancement for License Plate Recognition based on Deep Learning [C] . 2022 . |
| MLA | Wang, Kaixin et al. "Night Vision Enhancement for License Plate Recognition based on Deep Learning" . (2022) . |
| APA | Wang, Kaixin , Wu, Linhuang , Zhang, Shihao , Wen, Renfang . Night Vision Enhancement for License Plate Recognition based on Deep Learning . (2022) . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
Image deblurring refers to the deconvolution of blurry images and blur kernels to get sharp images.The process of deblurring when the blur kernel is unknown is called blind deblurring. In general, blind image deblurring includes two steps. One is the estimation of blur kernel, which is actually estimating how images become blurry. The other is to use the blur kernel for image deconvolution. In this paper, a blind deblurring method for natural images is proposed. Firstly, we take a preprocession and denoise the image to eliminate the influence of noise on kernel estimation and avoid the noise being amplified in the final image reconstruction stage. Secondly, we extract the main structure of the image by a relative total variation method, next enhance the structure image by the shock filter, and then extract its high frequency layer. This method can provide effective and significant edges for kernel estimation. In the kernel estimation stage, we use both the image intensity and gradient value as regularization terms to ensure the sparsity and continuity of the blur kernel. Finally, in image reconstruction stage, we use a simple hyper-Laplacian prior, which is fast and robust for small kernel errors. Extensive experiments showed that our method has the highest quality score quantitatively and the best visual effect qualitatively. © 2022 IEEE.
Keyword ：
Image denoising Image denoising Image enhancement Image enhancement Image reconstruction Image reconstruction
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | Wen, Renfang , Chen, Jian , Wu, Linhuang et al. A Robust Blind Deblurring Method for Natural Blurry Images [C] . 2022 . |
| MLA | Wen, Renfang et al. "A Robust Blind Deblurring Method for Natural Blurry Images" . (2022) . |
| APA | Wen, Renfang , Chen, Jian , Wu, Linhuang , Wang, Kaixin . A Robust Blind Deblurring Method for Natural Blurry Images . (2022) . |
| Export to | NoteExpress RIS BibTex |
Version ：
Abstract ：
为了提高视频编码的率失真性能,x265编码器中的前瞻模块为码率控制部分做了许多决策,如帧类型决策、编码单元CU的量化参数(Quantized Parameter,QP)偏移等.然而,x265中忽略了初始帧基础量化参数PQ0的确定,导致所有后续以初始帧为参考帧的帧图像质量下降.同时,帧级QP只由长期历史帧复杂度决定,导致对最近的帧变化不敏感.针对上述问题,提出一种充分利用前瞻模块来修正初始帧基础PQ0以及通过下采样SATD值改进P帧基础QP的算法.实验结果表明,在默认预设条件下,提出的算法使得率失真性能BD-rate(PSNR)与BD-rate(SSIM)分别降低了14.33％与16.22％,同时满足比特约束要求,有效提升了编码质量.
Keyword ：
x265 x265 前瞻模块 前瞻模块 码率控制 码率控制 视频编码 视频编码
Cite：
Copy from the list or Export to your reference management。
| GB/T 7714 | 陈华炜 , 吴林煌 . 一种基于x265的帧级基础量化参数确定方法 [J]. | 电视技术 , 2022 , 46 (12) : 77-82,88 . |
| MLA | 陈华炜 et al. "一种基于x265的帧级基础量化参数确定方法" . | 电视技术 46 . 12 (2022) : 77-82,88 . |
| APA | 陈华炜 , 吴林煌 . 一种基于x265的帧级基础量化参数确定方法 . | 电视技术 , 2022 , 46 (12) , 77-82,88 . |
| Export to | NoteExpress RIS BibTex |
Version ：
Export
| Results： |
Selected to |
| Format： |
