Lei Sun

Dr. Lei Sun


Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
19 April 2024 in Electronics
Verified
Journal Article
Open access
PDF file
YWYali Wu
Yali Wu
LSLei Sun
Lei Sun
XLXiaoshuang Liu
Xiaoshuang Liu
XLXiaoran Lin
Xiaoran Lin
As a contender in the competitive landscape of next-generation wireless communication technologies, visible light communication (VLC) stands out due to its potential for enhancing transmission rates and spectrum resource utilization. VLC offers various advantages, including license-free operation, high confidentiality, and cost-effectiveness. However, practical implementation faces challenges stemming from the limited modulation bandwidth of light-emitting diodes (LEDs), constraining system capacity and VLC communication rates. To address this limitation, non-orthogonal multiple access (NOMA) emerges as a novel multiple access strategy, particularly suitable for enhancing the capacity and communication rates of downlink VLC systems through power multiplexing. This paper delves into the energy-efficient design of joint LED association and power allocation (LA–PA) for downlink NOMA-based VLC systems. Through an analysis of channel capacity, we transform the non-convex energy-efficient optimization model, accounting for signal non-negativity, per-LED optical power constraints, and user rate constraints, into a convex form. Subsequently, we propose an iterative power allocation algorithm to attain solutions for the optimization problem with pre-established LED associations. Furthermore, we derive a feasibility condition for an LED association, considering signal non-negativity, per-LED optical power constraints, power constraints for successive interference cancellation (SIC), and channel gain between transceiver signals. This condition identifies feasible LEDs capable of maximizing energy efficiency (EE) when combined with the aforementioned power allocation algorithm. Finally, we illustrate the superiority of the joint LA–PA scheme in terms of the EE, transmission reliability, and transmission capacity performance gain over NOMA in the context of VLC.

2 January 2025 in Signal, Image and Video Processing
Verified
Journal Article
Closed access
PDF file
XMXiaoyue Ma
Xiaoyue Ma
LSLei Sun
Lei Sun
JGJieping Gao
Jieping Gao
YDYangming Dong
Yangming Dong

4 November 2025 in Research Square
Verified
Preprint
Open access
PDF file
JGJieping Gao
Jieping Gao
LSLei Sun
Lei Sun
PZPengyun Zhang
Pengyun Zhang
Skin cancer is one of the most common malignancies worldwide. The deep learning algorithms, such as convolutional neural networks (CNNs), plays a crucialrole in the automation of skin image analysis. However, the high complexityand large parameter sizes of CNNs limit their deployment on mobile or embedded systems. To address this problem, we propose Skin-LiteNet, a lightweightCNN model designed for efficient skin lesion classification. To enhance the feature extraction, the model use a multi branch DBConv module. Moreover, themodel introduces an ECSA attention mechanism to focus on critical lesion regionswhile maintaining low computational overhead. Experiments on the HAM10000dataset show that Skin-LiteNet achieves an accuracy of 98.97% and an F1 scoreof 98.89%, using only 284.44K parameters and 48.11M Flops. Compared withexisting models such as ResNet, MobileNetV2, ConvNeXt, and ShuffleNetV2,Skin-LiteNet offers better accuracy with reducing resource requirements significantly. These results indicate a strong potential for Skin-LiteNet in real-time andresource-constrained clinical applications.