Jinshu Huang

Jinshu Huang

Auto-generated profile, based on publicly available information

Jinshu Huang

Nanyang Institute of Technology

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 May 2025 in Journal of Electronic Imaging
Verified
Journal Article
Closed access
PDF file
TPTingxiao Pan
Tingxiao Pan
MHMing Hui
Ming Hui
JHJinshu Huang
Jinshu Huang
ZFZufeng Fu
Zufeng Fu
THTao Hai
Tao Hai
JYJinghong Yao
Jinghong Yao
With the rapid development of smart cities, vehicle detection algorithms have become a crucial component of intelligent transportation systems, widely applied in areas such as parking detection, traffic flow analysis, and illegal intrusion detection. However, due to limitations in computational resources and the increasing complexity of traffic scenarios, vehicle detection algorithms face considerable performance requirements and a demand for lightweight models. Therefore, we propose a lightweight vehicle detection model that incorporates a parallel multi-scale aggregation (PMSA) module into an improved You Only Look Once (YOLO) v5 architecture, termed PMSA-YOLO. First, the backbone network is optimized by designing a PMSA, which not only captures multi-scale information but also reduces reliance on network depth and width while maintaining recognition efficiency. Second, a convolutional block attention module is added to the last layer of the backbone network to enhance the extraction of key features, thereby improving feature recognition efficiency. Finally, a bidirectional feature pyramid network module is integrated into the neck network, and a more streamlined PMSA module is adopted for feature fusion. Experimental results demonstrate that relative to the original model, the improved model reduces the parameter count and computational load by 29.7% and 32.9%, respectively, while increasing precision, recall, mAP50, and mAP50–95 by 4.5%, 3.4%, 4.4%, and 2.6%, respectively.

5 June 2026 in Measurement Science and Technology
Verified
Journal Article
Closed access
ZJZhengju Jia
Zhengju Jia
MHMing Hui
Ming Hui
JHJinshu Huang
Jinshu Huang
ZFZufeng Fu
Zufeng Fu
TPTingxiao Pan
Tingxiao Pan
JYJinghong Yao
Jinghong Yao
Multispectral object detection leverages the complementary characteristics of visible (RGB) and infrared (IR) modalities to improve target perception under low illumination, occlusion, and other complex environmental conditions. However, existing methods often suffer from imbalanced cross-modal fusion, leading to incomplete information integration and insufficient exploitation of fine-grained complementary cues. To address these limitations, we propose FMAF-YOLO, a lightweight dual-branch network for multispectral object detection. In the backbone, a C3k2-MLCA module preserves coarse spatial structure via local average pooling and introduces a global–local dual-branch attention mechanism to enhance the representation of small objects and edge regions. In the neck, a multi-scale cross-modal fusion module performs adaptive inter-layer interaction and weighted fusion of RGB and IR features to improve fusion balance and complementarity. In addition, a channel-aware dynamic sampling module jointly models spatial and channel information during upsampling, reducing feature blurring and preserving fine-grained details. Experiments on two public multispectral benchmarks demonstrate the effectiveness of the proposed method. On the M 3 FD dataset, FMAF-YOLO improves mAP@0.5 and mAP@0.5:0.95 by 4.1 and 3.7 percentage points, respectively, over the YOLO11n baseline. On the DroneVehicle dataset, evaluated under the official public train/validation/test protocol, the corresponding gains are 2.7 and 2.4 percentage points. These results demonstrate that FMAF-YOLO achieves effective fine-grained feature representation and adaptive cross-modal fusion while maintaining competitive computational efficiency.