A deep learning based approach for automated plant disease classification using vision transformer

Borhani, Yasamin; Khoramdel, Javad; Najafi, Esmaeil

doi:10.1038/s41598-022-15163-0

Cited by 99 publications

(34 citation statements)

References 26 publications

(27 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many CNN models have already been proposed for image-based crop disease prediction and classification. The proposed work in [11] introduced a deep learning model based on weekly supervised learning. The work includes training the model on onion crops with six categories of symptoms and using the activation map to localize diseases.…”

Section: Related Workmentioning

confidence: 99%

A Convolutional Neural Network Model for Wheat Crop Disease Prediction

Ashraf¹,

Abrar²,

Qadeer³

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

Wheat is the most important cereal crop, and its low production incurs import pressure on the economy. It fulfills a significant portion of the daily energy requirements of the human body. The wheat disease is one of the major factors that result in low production and negatively affects the national economy. Thus, timely detection of wheat diseases is necessary for improving production. The CNN-based architectures showed tremendous achievement in the image-based classification and prediction of crop diseases. However, these models are computationally expensive and need a large amount of training data. In this research, a light weighted modified CNN architecture is proposed that uses eight layers particularly, three convolutional layers, three SoftMax layers, and two flattened layers, to detect wheat diseases effectively. The high-resolution images were collected from the fields in Azad Kashmir (Pakistan) and manually annotated by three human experts. The convolutional layers use 16, 32, and 64 filters. Every filter uses a 3 × 3 kernel size. The strides for all convolutional layers are set to 1. In this research, three different variants of datasets are used. These variants S1-70%:15%:15%, S2-75%:15%:10%, and S3-80%:10%:10% (train: validation: test) are used to evaluate the performance of the proposed model. The extensive experiments revealed that the S3 performed better than S1 and S2 datasets with 93% accuracy. The experiment also concludes that a more extensive training set with high-resolution images can detect wheat diseases more accurately.

show abstract

Section: Related Workmentioning

confidence: 99%

A Convolutional Neural Network Model for Wheat Crop Disease Prediction

Ashraf¹,

Abrar²,

Qadeer³

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

show abstract

“…The fine‐tuning process, which is applied similarly to its operation, is the process of increasing the image resolution value accepted to the network. The point of departure from this study is that instead of preparing images with different resolutions, the resolution value accepted by the network at the input has been changed 58 . Since there are images of different resolutions in our training dataset, blackening is applied without making any sense by the amount of missing pixels so that low‐resolution images can be entered into the high‐resolution network.…”

Section: Experimental Workmentioning

confidence: 99%

“…The point of departure from this study is that instead of preparing images with different resolutions, the resolution value accepted by the network at the input has been changed. 58 Since there are images of different resolutions in our training dataset, blackening is applied without making any sense by the amount of missing pixels so that low-resolution images can be entered into the high-resolution network. In other words, the applied fine-tuning is successful only in these images, as it does not play on the high-resolution images that are less in the data set.…”

Section: Comparison Metricsmentioning

confidence: 99%

Transfer learning and fine‐tuned transfer learning methods' effectiveness analyse in the CNN‐based deep learning models

Öztürk

Taşyürek

Türkdamar

2022

Concurrency and Computation

View full text Add to dashboard Cite

Summary Object detection is a type of application that includes computer vision and image processing technologies, which deal with detecting, tracking, and classifying desired objects in images. Computer vision is a field of artificial intelligence that enables computers and systems to derive information from digital images and take action or suggestions based on that information. CNN is one of the current methods of object detection due to its ease of use and GPU‐supported parallel working features. Due to the aim of completing deep learning model training quickly or due to insufficient dataset, many studies using the transfer learning method are carried out in fields such as medicine, agriculture, and weapons. However, there are very few studies that use the fine‐tuning method and compare transfer learning in terms of effectiveness. By paying attention to the balanced distribution of the data, approximately 100 images of each chess piece type were included in the analysis and a dataset of at least 1000 images was created. The without transfer learning fine‐tune, fine‐tuned transfer learning, transfer learning, fully supervised learning (FSL) and weakly supervised learning (WSL) applied models performances compared. Experimental results show that the fine‐tuned transfer learning applied YOLO V4 model produces more accurate results than the other models in FSL and the transfer learning applied Faster R‐CNN model produces more accurate results than the other models in WSL.

show abstract

“…Similar to CNN, which was previously applied to natural language processing, Vision Transformer (ViT) is one of the latest DL methods. One example is using ViT for plant disease classification . The authors compared the image recognition results of three models: (1) ViT, (2) CNN, and (3) the hybrid model of ViT and CNN under different dataset sizes.…”

Section: Introductionmentioning

confidence: 99%

Deep Learning–Assisted Surface-Enhanced Raman Scattering for Rapid Bacterial Identification

Tseng

Chen

Chao

et al. 2023

ACS Appl. Mater. Interfaces

View full text Add to dashboard Cite

Bloodstream infection (BSI) is characterized by the presence of viable microorganisms in the bloodstream and may induce systemic immune responses. Early and appropriate antibiotic usage is crucial to effectively treating BSI. However, conventional culture-based microbiological diagnostics are timeconsuming and cannot provide timely bacterial identification for subsequent antimicrobial susceptibility test (AST) and clinical decision-making. To address this issue, modern microbiological diagnostics have been developed, such as surface-enhanced Raman scattering (SERS), which is a sensitive, label-free, and quick bacterial detection method measuring specific bacterial metabolites. In this study, we aim to integrate a new deep learning (DL) method, Vision Transformer (ViT), with bacterial SERS spectral analysis to build the SERS-DL model for rapid identification of Gram type, species, and resistant strains. To demonstrate the feasibility of our approach, we used 11,774 SERS spectra obtained directly from eight common bacterial species in clinical blood samples without artificial introduction as the training dataset for the SERS-DL model. Our results showed that ViT achieved excellent identification accuracy of 99.30% for Gram type and 97.56% for species. Moreover, we employed transfer learning by using the Gram-positive species identifier as a pre-trained model to perform the antibiotic-resistant strain task. The identification accuracy of methicillin-resistant and -susceptible Staphylococcus aureus (MRSA and MSSA) can reach 98.5% with only 200-dataset requirement. In summary, our SERS-DL model has great potential to provide a quick clinical reference to determine the bacterial Gram type, species, and even resistant strains, which can guide early antibiotic usage in BSI.

show abstract

A deep learning based approach for automated plant disease classification using vision transformer

Cited by 99 publications

References 26 publications

A Convolutional Neural Network Model for Wheat Crop Disease Prediction

A Convolutional Neural Network Model for Wheat Crop Disease Prediction

Transfer learning and fine‐tuned transfer learning methods' effectiveness analyse in the CNN‐based deep learning models

Deep Learning–Assisted Surface-Enhanced Raman Scattering for Rapid Bacterial Identification

Contact Info

Product

Resources

About