Identification and Classification of Rice Plant Disease Using Hybrid Transfer Learning

Tunio, Muhammad Hanif; Li, Minglu; Butt, Muhammad Hassaan Farooq; Memon, Imran

doi:10.1109/iccwamtip53232.2021.9674124

Cited by 19 publications

(4 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These studies contributed to accurate classification of plant diseases. In addition, there are recent and strong models proposed for an accurate detection and classification like [21], [22]. Nevertheless, most of them focus on modifying the technique to achieve high accuracy and do not consider the effect of a number of parameters on deploying onto mobile applications.…”

Section: Related Workmentioning

confidence: 99%

“…Following that the CNN models have been improved time by time and achieved state-of-the-art results, particularly the agricultural sector [4]. However, these deep architectures and systems are time-consuming and little compatible with a low-level energy consumption architectures [22]. Therefore, proposing an accurate model while adapting well with time constraint, is an open question in literature up to this time.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

From a Proposed CNN Model to a Real-World Application in Rice Disease Classification

Nguyễn¹,

Tran²,

Thi³

et al. 2022

Proceedings of the Seventh International Conference on Research in Intelligent and Computing in Engineering

View full text Add to dashboard Cite

A compact and precise application of rice disease classification is helpful to assist farmers in their work for treatment on the plants and therefore could be quick and accurate to measure and eliminate the effects of diseases more profitably. In the past, the works were completed by naked-eye observation and basically relied on the experiences. Even so, the results are quite subjective and heuristic. In this paper, a mobile application to automatically classify several kinds of rice diseases from rice plant images and then to accurately recommend the uses of pesticides or chemicals. To do so, a proposed convolutional neural network (CNN) model is given. The results show that the proposed CNN model achieves the performance with the best trade-off between accuracy and time efficiency in comparison with the state-of-the-art models in our dataset. This model could be easily embedded into a mobile application to process in near real-time processing.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

From a Proposed CNN Model to a Real-World Application in Rice Disease Classification

Nguyễn¹,

Tran²,

Thi³

et al. 2022

Proceedings of the Seventh International Conference on Research in Intelligent and Computing in Engineering

View full text Add to dashboard Cite

show abstract

“…In order to solve the above problems, some scholars have conducted research based on deep learning and machine learning and achieved some results. Muhammad et al [3] combined deep convolutional neural network (CNN) and transfer learning to identify various rice diseases, with an accuracy of 80.8%. Rustia et al [4] proposed an automatic greenhouse pests and diseases identification algorithm based on the cascaded deep-learning classification method.…”

Section: Introductionmentioning

confidence: 99%

Design of greenhouse vegetable pest and disease identification method based on improved AlexNet model

Tang,

Aridas,

Talip

et al. 2024

Preprint

View full text Add to dashboard Cite

In China, many greenhouse farmers still rely on manual identification of vegetable pests and diseases. This method relies on traditional experience and intuitive observation, lacks scientific and systematic methods, which is prone to overuse in subsequent use of chemical pesticides. To this end, this study proposes a method for identifying greenhouse vegetable pests and diseases based on the improved AlexNet model. It uses the AlexNet as an image recognition model for pests and diseases and uses the ReLU6 activation function to solve the problems of poor model convergence and overfitting. It also integrates the GoogleNet Inception-v3 module to improve recognition results, which solves some problems of the AlexNet model, such as noise, poor model convergence, and over-fitting in target positioning. After compare with AlexNet, CNN (Convolutional Neural Networks), and YOLO-V3 (You Only Look Once Version 3.0) model, the IM-AlexNet model is superior to the other three models in MAP value, recognition accuracy, and loss function. It shows that the monitoring network designed in this study can better identify vegetable pests and diseases efficiently. It can help vegetable greenhouse farmers accurately and quickly identify vegetable pests and diseases, reduce the use of broad-spectrum pesticides, and save time and resources, which is beneficial to the environment and consumer health.

show abstract

“…Content retrieval (CR) is a procedure of acquiring and exploring certain data resources that are associated with some details of the resource pool (Mayr et al, 2018). In general, data is produced by different applications such as digital health records, web exploration, electronic libraries, and so on (Tunio et al, 2021). Since data resources generated after search process tend to vary in terms of supremacy and quantity, it is important to rate the information so as to describe the degree of significance (Deng & Liu, 2018; Singh, 2018).…”

Section: Introductionmentioning

confidence: 99%

Natural language processing with deep learning enabled hybrid content retrieval model for digital library management

et al. 2022

View full text Add to dashboard Cite

In recent times, natural language processing (NLP) technique has received significant attention in content retrieval (CR) domain. The emergence of digital libraries, in recent years, enables people from across the globe to access and store books, documents, and literature of multiple kinds. The development of NLP models has considerably improved the performance in terms of digital library management. In this scenario, artificial intelligence‐based expert systems are required to handle massive quantities of data that exist in digital libraries and achieve effective CR performance. In this background, the current study designs NLP with deep learning enabled hybrid content retrieval (NLPDL‐HCR) model for digital library management. The aim of the presented NLPDL‐HCR is to effectually retrieve the images as well as textual data from digital libraries based on a user's query. The proposed NLPDL‐HCR model encompasses two major stages namely, text retrieval and image retrieval (IR). During text retrieval process, the proposed NLPDL‐HCR model includes term frequency inverse document frequency vectorizer with optimal gated recurrent unit (GRU) model. The hyperparameters of the GRU model are optimally adjusted with the help of RMSProp approach. Besides, the IR process involves three sub‐processes namely, densely connected networks‐based feature extraction, butterfly optimization algorithm‐based hyperparameter tuning, and Euclidean distance‐based similarity measurement. The experimental analysis results, accomplished by the proposed NLPDL‐HCR model using benchmark datasets, highlighted its superior performance over recent state‐of‐the‐art approaches.

show abstract

Identification and Classification of Rice Plant Disease Using Hybrid Transfer Learning

Cited by 19 publications

References 8 publications

From a Proposed CNN Model to a Real-World Application in Rice Disease Classification

From a Proposed CNN Model to a Real-World Application in Rice Disease Classification

Design of greenhouse vegetable pest and disease identification method based on improved AlexNet model

Natural language processing with deep learning enabled hybrid content retrieval model for digital library management

Contact Info

Product

Resources

About