Linjing Wei

Prof. Dr. Linjing Wei

orcid

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
21 July 2024 in Agronomy
Verified
Journal Article
Open access
PDF file
ZJZhengjie Ji
Zhengjie Ji
SBShudi Bao
Shudi Bao
MCMeng Chen
Meng Chen
LWLinjing Wei
Linjing Wei
The accurate identification of corn leaf diseases is crucial for preventing disease spread and improving corn yield. Plant leaf images are often affected by factors such as complex backgrounds, climate, light, and sample data imbalance. To address these issues, we propose a lightweight convolutional neural network, ICS-ResNet, based on ResNet50. This network incorporates improved spatial and channel attention modules as well as a deep separable residual structure to enhance recognition accuracy. (1) The residual connections in the ResNet network prevent gradient loss during deep network training. (2) The improved channel attention (ICA) and spatial attention (ISA) modules fully utilize semantic information from different feature layers to accurately localize key features of the network. (3) To reduce the number of parameters and lower computational costs, we replace traditional convolutional computation with a depth-separable residual structure. (4) We also employ cosine annealing to dynamically adjust the learning rate, enhancing the network’s training stability, improving model convergence, and preventing local optima. Experiments on the corn dataset in Plant Village compare the proposed ICS-ResNet with eight popular networks: CSPNet, InceptionNet_v3, EfficientNet, ShuffleNet, MobileNet, ResNet50, ResNet101 and ResNet152. The results show that the ICS-ResNet achieves an accuracy of 98.87%, which is 5.03%, 3.18%, 1.13%, 1.81%, 1.13%, 0.68%, 0.44% and 0.60% higher than the other networks, respectively. Furthermore, the number of parameters and computations are reduced by 69.21% and 54.88%, respectively, compared to the original ResNet50 network, significantly improving the efficiency of corn leaf disease classification. The study provides strong technical support for sustainable agriculture and the promotion of agricultural science and technology innovation.

15 October 2025 in Scientific Reports
Verified
Journal Article
Open access
PDF file
TZTianxiang Zhou
Tianxiang Zhou
LWLinjing Wei
Linjing Wei
During rice cultivation, common rice diseases and pests such as Rice blast, Bacterial blight, Brown-planthopper and Leaf-folder will significantly affect the yield and quality. The current model is limited to detecting rice diseases or pests alone, and faces challenges such as the diversity of disease and pest traits, small detection targets, uneven light and complex background shading in paddy fields, resulting in low accuracy and adaptability of the model. In this study, a YOLO-DP (Diseases and Pests) model based on YOLOv8n model was proposed to detect fifteen common rice diseases and pests under complex conditions. First, the Triplet Attention mechanism is introduced into the network’s Backbone to achieve cross-dimensional interaction between channels and spatial dimensions. Then, GLSA (the Global to Local Spatial Aggregation) module is used to improve the Neck of YOLOv8n, enhancing the effectiveness of feature representation. The WTConv (Wavelet Transform Convolution) is used to improve the C2f-BottleNeck of the original model, expanding the network’s receptive field. Finally, the loss function is replaced with EIoU (Enhanced Intersection over Union) to reduce the position offset and shape mismatch of the predicted boxes. Experimental results demonstrate that the model achieves an average accuracy of 80.9%, a recall rate of 74.4%, a Mean Average Precision mAP50 of 77.8% and mAP95 of 50.1%, significantly outperforming the original YOLOv8n and mainstream detection models such as TOOD, Faster R-CNN and RT-DETR. This model exhibits exceptional performance in detecting rice diseases and pests in complex environments, providing robust technical support for rice growth monitoring and offering insights for the detection of other crop diseases.

20 May 2025 in Journal of King Saud University Computer and Information Sciences
Verified
Journal Article
Open access
PDF file
HMHenghui Mo
Henghui Mo
LWLinjing Wei
Linjing Wei
In cotton cultivation, diseases such as leaf spot, helminthosporium leaf spot, fusarium wilt, boll gray mold, and leaf curl significantly affect yield and quality. Current models face challenges like diverse disease traits, variable stages, small target detection, uneven lighting, and occlusions, resulting in low accuracy and adaptability. This study introduces the RT-DETR-DFSA model for cotton leaf disease detection, building on the Real-Time Detection Transformer (RT-DETR). The Decoupled Focused Self-Attention (DFSA) mechanism splits traditional two-dimensional self-attention into one-dimensional operations that are processed by a dilated convolution layer, merges positional features with the original input, enhances feature relationships, and dynamically adjusts self-attention weights. Style Generative Adversarial Network with Adaptive Discriminator Augmentation (StyleGAN2-ADA) and Fourier Transform are used to generate realistic images of cotton diseases, enhancing training and validation sets. The Layer-Adaptive Magnitude-based Pruning (LAMP) method reduces computational and memory demands, and the teacher-assistant-student architecture further improves accuracy through knowledge distillation. The RT-DETR-DFSA model achieves 87.14% detection accuracy, 85.03% recall, and 86.33% mean Average Precision (mAP50). Post-pruning, the model’s parameters are reduced to 4.9 million (M), with a computational demand of 31.5 Giga Floating-Point Operations Per Second (GFLOPs), showing superior performance over existing models. This provides technical support for cotton monitoring and insights for disease detection in other crops.

5 August 2025 in Frontiers in Education
Verified
Journal Article
Open access
PDF file
MZMeiqi Zhong
Meiqi Zhong
LWLinjing Wei
Linjing Wei
HMHenghui Mo
Henghui Mo
Introduction: With the rapid advancement of artificial intelligence (AI) technologies, there is a growing demand in graduate education for highly skilled, application-oriented professionals equipped with both technical expertise and strong ethical values. However, traditional AI curricula often lack effective integration of practical training and values-based education.Methods: This study describes the design and implementation of an integrated curriculum for AI master's students, combining systematic industry needs assessment, curricular reform, and pedagogical innovation. The program emphasizes the cultivation of practical skills alongside values-oriented education—specifically professional responsibility, ethical awareness, and societal engagement—through course planning, case-based learning, project-driven instruction, and interdisciplinary collaboration. Multiple evaluation tools were employed, including student and faculty surveys, focus group interviews, and analysis of performance metrics such as project outcomes and academic competition results.Results: Evaluation results demonstrate that the integrated curriculum significantly enhances students' initiative, critical thinking ability, and sense of social responsibility. Students not only achieved higher technical competence but also exhibited improved holistic qualities, as reflected in project achievements and external competition performance.Discussion: The core innovation of this model lies in the seamless integration of ethical and societal considerations into technical training. Although the case study is limited to a single institution, the model provides valuable insights for AI education reform and high-quality talent cultivation. These findings establish a theoretical foundation for the adaptation and application of such integrative approaches across diverse educational settings.

9 July 2024 in Mathematics
Verified
Journal Article
Open access
PDF file
HMHenghui Mo
Henghui Mo
LWLinjing Wei
Linjing Wei
In response to the current lack of annotations for flower images and insufficient focus on key image features in traditional fine-grained flower image classification based on deep learning, this study proposes the SA-ConvNeXt flower image classification model. Initially, in the image preprocessing stage, a padding algorithm was used to prevent image deformation and loss of detail caused by scaling. Subsequently, the model was integrated using multi-level feature extraction within the Efficient Channel Attention (ECA) mechanism, forming an M-ECA structure to capture channel features at different levels; a pixel attention mechanism was also introduced to filter out irrelevant or noisy information in the images. Following this, a parameter-free attention module (SimAM) was introduced after deep convolution in the ConvNeXt Block to reweight the input features. SANet, which combines M-ECA and pixel attention mechanisms, was employed at the end of the module to further enhance the model’s dynamic extraction capability of channel and pixel features. Considering the model’s generalization capability, transfer learning was utilized to migrate the pretrained weights of ConvNeXt on the ImageNet dataset to the SA-ConvNeXt model. During training, the Focal Loss function and the Adam optimizer were used to address sample imbalance and reduce gradient fluctuations, thereby enhancing training stability. Finally, the Grad-CAM++ technique was used to generate heatmaps of classification predictions, facilitating the visualization of effective features and deepening the understanding of the model’s focus areas. Comparative experiments were conducted on the Oxford Flowers102 flower image dataset. Compared to existing flower image classification technologies, SA-ConvNeXt performed excellently, achieving a high accuracy of 96.7% and a recall rate of 98.2%, with improvements of 4.0% and 3.7%, respectively, compared to the original ConvNeXt. The results demonstrate that SA-ConvNeXt can effectively capture more accurate key features of flower images, providing an effective technical means for flower recognition and classification.

4 September 2025 in Frontiers in Plant Science
Verified
Journal Article
Open access
PDF file
XZXin Zhang
Xin Zhang
LWLinjing Wei
Linjing Wei
RYRuqiang Yang
Ruqiang Yang
This study proposes EDGE-MSE-YOLOv11, a novel lightweight rice disease detection model based on a unified Tri-Module Lightweight Perception Mechanism (TMLPM). This mechanism integrates three core components: multi-scale feature fusion (C3K2 MSEIE), attention-guided feature refinement (SimAM), and efficient spatial downsampling (ADown), which significantly enhance the model’s ability to detect multi-scale and small disease targets under complex field conditions. Unlike isolated architectural enhancements, TMLPM supports collaborative feature interactions, which significantly improves the interpretability and computational efficiency of the model under complex environmental conditions. Experimental results show that, compared with the baseline YOLOv11n model, EDGE-MSE-YOLOv11 improves precision (from 85.6% to 89.2%), recall (from 82.6% to 86.4%), mAP@0.5 (from 90.2% to 92.6%), and mAP@0.5:0.95 (from 63.7% to 70.3%). The model also reduces parameter count by 0.69M and computational cost by 0.3 GFLOPs, while maintaining a high inference speed of 111.6 FPS. These results validate its effectiveness in identifying small, dense lesion areas with high accuracy and efficiency. However, the model still faces challenges in detecting ultra-small or occluded lesions under extremely complex conditions and has yet to be evaluated across multiple domains. Future work will focus on cross-domain generalization and deployment optimization using lightweight techniques such as quantization, pruning, and transformer-based enhancements, aiming to build a robust and scalable disease diagnosis system for intelligent agriculture.