Optimized threshold-based convolutional neural network for plant leaf classification: a challenge towards untrained data

Dudi, Bhanuprakash; Rajesh, V.

doi:10.1007/s10878-021-00770-w

Cited by 17 publications

(9 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The population size and the maximum iterations were taken as 10 and 25. The introduced BM‐HHO‐CRNN + CSVM was compared with “various existing heuristic‐based algorithms like HHO‐CRNN + CSVM (Heidari et al, 2019), Shark Smell‐based Whale Optimization Algorithm (SS‐WOA)‐CRNN + CSVM (Dudi & Rajesh, 2021), WOA‐CRNN + CSVM (Mirjalili & Lewis, 2016), and Shark Smell Optimization (SSO)‐CRNN + CSVM (Abedinia et al, 2014) and with different conventional classifiers like SVM (Yu et al, 2016), RNN (Dudi & Rajesh, 2021), CNN (Rawat & Wang, 2017), CSVM (Rawat & Wang, 2017; Yu et al, 2016), CRNN (Rawat & Wang, 2017; Rong et al, 2014), and CRNN + CSVM (Rawat & Wang, 2017; Rong et al, 2014; Yu et al, 2016)” to prove the better performance offered by the introduced method for classifying the plant leaf from the trained and untrained data.…”

Section: Computation Of Resultsmentioning

confidence: 99%

Impact of modified Harris hawks optimization on hybrid deep learning for untrained plant leaf classification

Dudi,

Rajesh

2024

Journal of Field Robotics

Self Cite

View full text Add to dashboard Cite

Plants are extensively dispersed in nature, and they are essential for the survival and growth of the entire living creatures on the planet. Leaf classification and identification are critical in botany for recognizing new or endangered tree species. Leaf classification must be done quickly and accurately to sustain agricultural products. Traditional methods used manual classification of plant species. However, this method is costly, time‐consuming, and in some circumstances, impracticable. In the domains of image classification, target identification, and other fields, the new generation of Convolutional Neural Networks (CNNs) has shown outstanding outcomes. However, there exist some practical difficulties in executing these artificial neural networks. Owing to its high complexity and colossal running time, and also insufficient efficiency, the performance of classification is affected. An inappropriate count of neurons at each hidden layer may lead to an overfitting or underfitting problem. Thus, the core concept of this task is concerned with a novel technique for the growth of plant leaf classification for untrained data by hybridized classifiers. The core concept of this task is to find untrained data. The leaf images are collected from online sources. The images are further fed to the preprocessing phase. Pattern extraction is done on the preprocessed image using Weighted Weber Local Pattern. This preprocessed image is used for the classification of untrained data using hybridized deep‐structured architectures. Here, the combination of CNN with Recurrent Neural Network (RNN) and CNN with Support Vector Machine (SVM) is utilized for classification. Averaging the classification score of both hybrid techniques with parameter and threshold being optimized by the Best Searchable Modified Harris Hawks Optimization (BM‐HHO) produces the best classification outcome for untrained data. Here, the corresponding leaf type is classified if the input is given as the trained leaf images. Simultaneously, if the other types of data, like, palm tree leaf images, are given as input, and then explicitly shows the output as another type of leaves (or) untrained data. The experimental outcomes on the benchmark data set containing leaf images reveal that the introduced method achieved a higher accuracy than the conventional techniques. The accuracy of BM‐HHO‐CRNN + CSVM at the 65th learning rate is 3.43%, 0.84%, 1.80%, and 1.05% superior to HHO‐CRNN + CSVM, SS‐WOA‐CRNN + CSVM, WOA‐CRNN + CSVM, and SSO‐CRNN + CSVM.

show abstract

Section: Computation Of Resultsmentioning

confidence: 99%

Impact of modified Harris hawks optimization on hybrid deep learning for untrained plant leaf classification

Dudi,

Rajesh

2024

Journal of Field Robotics

Self Cite

View full text Add to dashboard Cite

show abstract

“…Dudi and Rajesh [40] proposed the shark smell-based whale optimization algorithm (SS-WOA), which identifies untrained class data based on the activation function values and classification costs of the CNN model. The method derives a threshold using the activation function value when data are input and identify the input data as untrained class data when the classification cost is less than the threshold.…”

Section: Related Workmentioning

confidence: 99%

“…The method improves the classification accuracy by approximately 0.57-4% compared with other classification models in the experiment on plant-leaf datasets. The methods proposed in [13] and [40] increase the computational cost because separate modules (extra layer and algorithm) are added to the CNN model.…”

Section: Related Workmentioning

confidence: 99%

“…Although data-based studies can identify data of untrained types at a lower cost than model-based studies, the accuracy is low. Model-based studies identify the data of new untrained types by adding prototypes [16,24] or adding or modifying extra structures or modules to the model [13,40]. Model-based studies increase the computational cost and require more resources owing to the increased number of parameters in the model, which can affect its performance.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Identification of untrained class data using neuron clusters

Lee

Chae

2023

Neural Comput & Applic

View full text Add to dashboard Cite

Convolutional neural networks (CNNs), a representative type of deep neural networks, are used in various fields. There are problems that should be solved to operate CNN in the real-world. In real-world operating environments, the CNN’s performance may be degraded due to data of untrained types, which limits its operability. In this study, we propose a method for identifying data of a type that the model has not trained on based on the neuron cluster, a set of neurons activated based on the type of input data. In experiments performed on the ResNet model with the MNIST, CIFAR-10, and STL-10 datasets, the proposed method identifies data of untrained and trained types with an accuracy of 85% or higher. The more data used for neuron cluster identification, the higher the accuracy; conversely, the more complex the dataset's characteristics, the lower the accuracy. The proposed method uses only the information of activated neurons without any addition or modification of the model’s structure; hence, the computational cost is low without affecting the classification performance of the model.

show abstract

“…The proposed method was more successful than AlexNet in the presence of less training data. Dudi and Rajesh [30] performed a DL-based plant identification application using the Swedish and Mendeley dataset. First, they applied pre-processing steps such as median filtering and histogram equalization to the leaves.…”

Section: Related Workmentioning

confidence: 99%

Comparative Analysis of CNN Models and Bayesian Optimization-Based Machine Learning Algorithms in Leaf Type Classification

Aslan

2023

Balkan Journal of Electrical and Computer Engineering

View full text Add to dashboard Cite

In this study, the leaves are classified by various Machine Learning (ML) and Deep Learning (DL) based Convolutional Neural Networks (CNN) methods. In the proposed method, first, image pre-processing is performed to increase the accuracy of the posterior process. The obtained image is a grayscale image without noise as a result of the pre-processing. These preprocessed images are used in classification with ML and DL. The Speeded Up Robust Features (SURF) are extracted from the grayscale image for ML-based learning. The features are restructured as visual words using the Bag of Visual Words (BoVW) method. Then, histograms are generated for each image according to the frequency of the visual word. Those histograms represent the new feature data. The histogram features are classified by four different ML methods, Decision Tree (DT), k-Nearest Neighbor (KNN), Naive Bayes (NB) and Support Vector Machine (SVM). Before using the ML methods, Bayesian Optimization (BO) method, which is one of the Hyperparameter Optimization (HO) algorithms, is applied to determine hyperparameters. In the classification process performed with four different ML algorithms, the best accuracy is achieved with the KNN algorithm as 98.09%. Resnet18, ResNet50, MobileNet, GoogLeNet, DenseNet, which are state-of-the-art CNN architectures, are used for DL-based learning. CNN models have higher accuracy than ML algorithms.

show abstract

Optimized threshold-based convolutional neural network for plant leaf classification: a challenge towards untrained data

Cited by 17 publications

References 36 publications

Impact of modified Harris hawks optimization on hybrid deep learning for untrained plant leaf classification

Impact of modified Harris hawks optimization on hybrid deep learning for untrained plant leaf classification

Identification of untrained class data using neuron clusters

Comparative Analysis of CNN Models and Bayesian Optimization-Based Machine Learning Algorithms in Leaf Type Classification

Contact Info

Product

Resources

About