A hybrid model of ghost-convolution enlightened transformer for effective diagnosis of grape leaf disease and pest

Lǚ, Xiangyu; Yang, Rui; Zhou, Jun; Jiao, Jie; Liu, Fei; Liu, Yuanyuan; Su, Baofeng; Pei-wen, GU

doi:10.1016/j.jksuci.2022.03.006

Cited by 31 publications

(15 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, the evaluation of hyperparameter L showed the accuracy and inference speed increased with the decrease of L , which might imply a saturation of the model. A similar result occurred in Lu et al's study ( Lu et al, 2022 ). Therefore, ResTFG (C13, H12, L2) achieved the best results with 10.86M number of parameters, 216FPS, and 97.1% accuracy.…”

Section: Resultssupporting

confidence: 89%

“…Appropriate data augmentation was conducted for categories of images with fewer samples. Specifically, all E. Maxima images were flipped horizontally, and 300 E. Brunetti images and 200 E. Necatrix images were randomly selected for horizontal flipping, which is a commonly used method for dataset augmentation ( Wang et al, 2019 ; Ye et al, 2020 ; Lu et al, 2022 ). Considering that the imaging environment of micrographs is controllable and consistent, and the original micrographs are sufficiently representative, it is not necessary to augment the original dataset by utilizing some morphological or color adjustment methods.…”

Section: Methodsmentioning

confidence: 99%

“…Transformer can capture global features well, but is less capable of capturing local features. Therefore, integrating CNN and Transformer structure could improve the model performance ( Dai et al, 2021 ; Lu et al, 2022 ).…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A reliable and low-cost deep learning model integrating convolutional neural network and transformer structure for fine-grained classification of chicken Eimeria species

He¹,

Chen²,

He³

et al. 2023

Poultry Science

View full text Add to dashboard Cite

Section: Resultssupporting

confidence: 89%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A reliable and low-cost deep learning model integrating convolutional neural network and transformer structure for fine-grained classification of chicken Eimeria species

He¹,

Chen²,

He³

et al. 2023

Poultry Science

View full text Add to dashboard Cite

“…The Ghost convolution module is divided into three parts: regular convolution [26], group convolution and identity operation. The first step of regular convolution uses a convolution filter set to generate part of the input feature map, then group convolution further uses a simple linear transformation operation on this part of the feature map, and finally the feature maps generated in the first two parts are stitched together by the identity constant operation to generate the final feature map, and the schematic diagram of the Ghost module is shown in Fig 2.…”

Section: A the C3ghost Modulementioning

confidence: 99%

Research on Defect Detection Method of Railway Transmission Line Insulators Based on GC-YOLO

Ding,

Rao,

Ding

et al. 2023

IEEE Access

View full text Add to dashboard Cite

The insulator defect targets in UAV (Unmanned Aerial Vehicle) images are often small and set against complex backgrounds. Consequently, traditional object detection algorithms commonly struggle to identify these minor defects. To enhance precision and recall in detecting insulator defects, a novel model, GC-YOLO (ghost convolution and centralized feature pyramid -You Only Look Once), based on YOLOv5s, has been introduced. GC-YOLO incorporates the Ghost convolution module in the backbone network, reducing feature redundancy and improving the inference speed of the feature extraction network. Moreover, an attention mechanism based on Coordinate Attention (CA) is integrated at the terminal of the backbone network, aimed at emphasizing the extraction of crucial information from target features. The Explicit Visual Center Block (EVCBlock) module from Centralized Feature Pyramid Network (CFPNet) is introduced in the neck layer to effectively fuse multi-scale features and enhance the feature map's characterization capability. Furthermore, in order to enhance the precision in detecting small-sized defects, a small object detection head is also added to the detection layer based on CFPNet. Experimental results demonstrate that GC-YOLO achieves a recall of 89.7% and

show abstract

“…Hybrid models combine the strengths of CNN and Transformer to improve performance. A model based on CNN called Ghost-Enlightened Transformer, for example, was proposed by [ 28 ] to construct intermediate feature maps. In the next step, the self-attention mechanism is used to convert those maps into deep semantic features.…”

Section: Introductionmentioning

confidence: 99%

Visual Intelligence in Precision Agriculture: Exploring Plant Disease Detection via Efficient Vision Transformers

Parez,

Dilshad,

Alghamdi

et al. 2023

Sensors

View full text Add to dashboard Cite

In order for a country’s economy to grow, agricultural development is essential. Plant diseases, however, severely hamper crop growth rate and quality. In the absence of domain experts and with low contrast information, accurate identification of these diseases is very challenging and time-consuming. This leads to an agricultural management system in need of a method for automatically detecting disease at an early stage. As a consequence of dimensionality reduction, CNN-based models use pooling layers, which results in the loss of vital information, including the precise location of the most prominent features. In response to these challenges, we propose a fine-tuned technique, GreenViT, for detecting plant infections and diseases based on Vision Transformers (ViTs). Similar to word embedding, we divide the input image into smaller blocks or patches and feed these to the ViT sequentially. Our approach leverages the strengths of ViTs in order to overcome the problems associated with CNN-based models. Experiments on widely used benchmark datasets were conducted to evaluate the proposed GreenViT performance. Based on the obtained experimental outcomes, the proposed technique outperforms state-of-the-art (SOTA) CNN models for detecting plant diseases.

show abstract

A hybrid model of ghost-convolution enlightened transformer for effective diagnosis of grape leaf disease and pest

Cited by 31 publications

References 36 publications

A reliable and low-cost deep learning model integrating convolutional neural network and transformer structure for fine-grained classification of chicken Eimeria species

A reliable and low-cost deep learning model integrating convolutional neural network and transformer structure for fine-grained classification of chicken Eimeria species

Research on Defect Detection Method of Railway Transmission Line Insulators Based on GC-YOLO

Visual Intelligence in Precision Agriculture: Exploring Plant Disease Detection via Efficient Vision Transformers

Contact Info

Product

Resources

About