A hybrid approach combining CUR matrix decomposition and weighted kernel sparse representation for plant leaf recognition

Plants play a crucial role in foodstuff, medicine, industry, and environmental protection. The skill of recognising plants is very important in some applications, including conservation of endangered species and rehabilitation of lands after mining activities. However, it is a difficult task to identify plant species because it requires specialized knowledge. Developing an automated classification system for plant species is necessary and valuable since it can help specialists as well as the public in identifying plant species easily. Shape descriptors were applied on the myDAUN dataset that contains 45 tropical shrub species collected from the University of Malaya (UM), Malaysia. Based on literature review, this is the first study in the development of tropical shrub species image dataset and classification using a hybrid of leaf shape and machine learning approach. Four types of shape descriptors were used in this study namely morphological shape descriptors (MSD), Histogram of Oriented Gradients (HOG), Hu invariant moments (Hu) and Zernike moments (ZM). Single descriptor, as well as the combination of hybrid descriptors were tested and compared. The tropical shrub species are classified using six different classifiers, which are artificial neural network (ANN), random forest (RF), support vector machine (SVM), k-nearest neighbour (k-NN), linear discriminant analysis (LDA) and directed acyclic graph multiclass least squares twin support vector machine (DAG MLSTSVM). In addition, three types of feature selection methods were tested in the myDAUN dataset, Relief, Correlation-based feature selection (CFS) and Pearson’s coefficient correlation (PCC). The well-known Flavia dataset and Swedish Leaf dataset were used as the validation dataset on the proposed methods. The results showed that the hybrid of all descriptors of ANN outperformed the other classifiers with an average classification accuracy of 98.23% for the myDAUN dataset, 95.25% for the Flavia dataset and 99.89% for the Swedish Leaf dataset. In addition, the Relief feature selection method achieved the highest classification accuracy of 98.13% after 80 (or 60%) of the original features were reduced, from 133 to 53 descriptors in the myDAUN dataset with the reduction in computational time. Subsequently, the hybridisation of four descriptors gave the best results compared to others. It is proven that the combination MSD and HOG were good enough for tropical shrubs species classification. Hu and ZM descriptors also improved the accuracy in tropical shrubs species classification in terms of invariant to translation, rotation and scale. ANN outperformed the others for tropical shrub species classification in this study. Feature selection methods can be used in the classification of tropical shrub species, as the comparable results could be obtained with the reduced descriptors and reduced in computational time and cost.

In this paper, the utility of data augmentation in plant recognition is presented. For this, the Swedish Leaf dataset is used. It covers 15 species with 75 images per species. After augmenting each of the images used for training by flipping, rotating, and random cropping, the total number of training images increased from 855 to 3420. The augmented dataset is then tested on a convolutional neural network developed from scratch. The test results showed 97.78% accuracy with the augmented dataset, against 88.52% on the original dataset.KeywordsSmart farmingPlant recognitionData augmentationConvolutional neural networkSwedish leaf dataset

Plant leaf classification plays a foremost role in botanical research, Ayurveda, agriculture practices, medicine and drug, weed detection, and many more areas. It is a technique by which a plant leaf is categorized based on its different morphological structures. The aim of this paper is to offer a deep learning technique for plant leaf classification with the help of deep Convolutional Neural Network as a substitute of conventional classification methods like k-nearest neighbor, probabilistic neural network, support vector machine, genetic algorithm, and principal component analysis, which all need feature extraction and are time-consuming. CNN has been used since it has outperformed in various image recognition challenges and feature extraction is performed automatically, which takes less time. This work uses a 5000 leaf images of two plant species. 4000 images are used for training and 1000 for testing purpose. CNN is trained in such a way that it can classify the species and predict the class for a new leaf image. The proposed model is run for different epochs and results were recorded. It was observed that the CNN model performed effectively in distinguishing the plant leaf images and achieved 99.96% training accuracy and 99.90% testing accuracy.

Detection of congestive heart failure from short-term heart rate variability segments using hybrid feature selection approach

Urbanization and expansion in each city of emerging countries have become an essential function of Earth’s surface, with the majority of people migrating from rural to urban regions. The various urban category characteristics have emphasized the great importance of understanding and creating suitable land evaluations in the future. The overall objective of this study is to classify the urban zone utilizing building height which is estimated using Sentinel-1 synthetic aperture radar (SAR) and various satellite-based indexes of Sentinel-2A. The first objective of this research is to estimate the building height of the Sentinel-1 SAR in Nonthaburi, Thailand. A new indicator, vertical-vertical-horizontal polarization (VVH), which can provide a better performance, is produced from the dual-polarization information, vertical-vertical (VV), and vertical-horizontal (VH). Then, the building height model was developed using indicator VVH and the reference building height data. The root means square error (RMSE) between the estimated and reference height is 1.413 m. Then, the second objective is to classify three classes of urban types, which are composed of residential buildings, commercial buildings, and other buildings, including vegetation, waterbodies, car parks, and so on. Spectral indices such as normalized difference vegetation index (NDVI), normalized difference water index (NDWI), and normalized difference built up the index (NDBI) are extracted from the Sentinel-2A data. To classify the urban types, a three-machine learning classifier, support vector machine (SVM), random forest (RF), and k-nearest neighbor (KNN) were developed. The classification uses randomly trained data from each 500 m focus study which are divided into a 100 × 100 m grid. Different models are examined using different variables, for example, classification using only building height and only spectral indices. The indices and estimated building height were used to classify the urban types. Not only the average of various satellite-based indices and building height of 100 × 100 m grid was used, but also the minimum, maximum, mean, and standard deviation were calculated from NDVI, NDWI, NDBI, and building height. There are a total of 16 variables used in the model. Eventually, the principal components analysis (PCA) was used to reduce the variables and get better performance of the models. SVM showed better accuracy than the other two, RF and KNN. The accuracies of SVM, RF, and KNN are 0.86, 0.75, and 0.76, respectively.

This paper carries on research surrounding the influences produced by dimensionality reduction on machine learning classification effect. Firstly, paper constructs the analysis architecture of data dimension reduction classification, combines the two different unsupervised dimension reduction methods, locally linear embedding (LLE) and principal component analysis (PCA) with the five machine learning classification methods: Gradient Boosting Decision Tree (GBDT), Random Forest, Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Logistic Regression. And then uses the handwritten digital identification dataset to analyze the classification performance of these five classification methods on different dimension datasets by different dimensionality reduction methods. The analysis shows that using the appropriate dimensionality reduction method for dimensionality reduction classification can effectively improve the classification accuracy; the dimensionality reduction classification effect of non-linear dimensionality reduction method is generally better than the linear dimensionality reduction method; different machine learning classification algorithms have significant differences in the sensitivity of dimensions.

In the absence of plants, we would not be able to live on this planet. A growing number of plant species are available, and they number in the hundreds. Foresters, farmers, environmentalists, and educators need knowledge of species to make informed decisions. As a result, species identification belongs to an interdisciplinary field of study. It may be challenging and tedious for nonexperts who are not familiar with typical botanical terms. In the fields of computer vision and machine learning, advances in this area can make this task easier. Although some efforts have been made, all plant species cannot be identified with one system. In this study, a similar approach was taken. In order to identify a plant, four steps have to be completed: acquisition of images, pre-processing, feature extraction, and classification. For this study, 1,125 images of 15 different species of leaves were used from the Swedish leaf dataset. Pre-processing is performed using Gaussian filtering mechanisms, and then texture and colour features are extracted.

A single nucleotide polymorphism (SNP) is a DNA sequence variation in a population. SNP is just a single nucleotide difference in the genome. Many statistical methods have been proposed to predict the racial classification of individuals based on SNP genetic data. The selection of the right classification method is very important because it will determine the accuracy of the classification results. This research aims to identify the highest average accuracy between two popular classification methods in Machine Learning (ML), including K-Nearest Neighborhood (KNN) and Support Vector Machine (SVM). This study used SNP genetic data for 120 samples from 2 CEU-European races and Yoruba-African races, where for each sample 10 SNPs were selected with the same location identity. The experiment was carried out by testing each classification method with variations in the percentage of test data 10, 20, 30, 40 and 50, which was combined with Euclidean distance for the KNN classification method. Based on the results of the study, the accuracy of the prediction of the classification of individual races based on SNP genetic data, the classification using KNN has an average prediction accuracy that is better than the SVM classification if the SNP location used tests has a high correlation with the sample class. In this case, the highest average accuracy value of KNN is 98.906% and SVM is 98.779%. There is a significant difference between the highest average accuracy of KNN and SVM based on the Wilcoxon statistical test with a significance level of α = 0.05. Benefits of this research are to find the right classification method for predictions of individual racial classification based on SNP genetic data.

Oesophagectomy is an operation with a high risk of postoperative complications. The aim of this single-centre retrospective study was to apply machine-learning methods to predict complications (Clavien-Dindo grade IIIa or higher) and specific adverse events. Patients with resectable adenocarcinoma or squamous cell carcinoma of the oesophagus and gastro-oesophageal junction who underwent Ivor Lewis oesophagectomy between 2016 and 2021 were included. The tested algorithms were logistic regression after recursive feature elimination, random forest, k-nearest neighbour, support vector machine, and neural network. The algorithms were also compared with a current risk score (the Cologne risk score). 457 patients had Clavien-Dindo grade IIIa or higher complications (52.9 per cent) versus 407 patients with Clavien-Dindo grade 0, I, or II complications (47.1 per cent). After 3-fold imputation and 3-fold cross-validation, the overall accuracies were: logistic regression after recursive feature elimination, 0.528; random forest, 0.535; k-nearest neighbour, 0.491; support vector machine, 0.511; neural network, 0.688; and Cologne risk score, 0.510. For medical complications, the results were: logistic regression after recursive feature elimination, 0.688; random forest, 0.664; k-nearest neighbour, 0.673; support vector machine, 0.681; neural network, 0.692; and Cologne risk score, 0.650. For surgical complications, the results were: logistic regression after recursive feature elimination, 0.621; random forest, 0.617; k-nearest neighbour, 0.620; support vector machine, 0.634; neural network, 0.667; and Cologne risk score, 0.624. The calculated area under the curve of the neural network was 0.672 for Clavien-Dindo grade IIIa or higher, 0.695 for medical complications, and 0.653 for surgical complications. The neural network scored the highest accuracies compared with all of the other models for the prediction of postoperative complications after oesophagectomy.

Detecting the presence of a disease requires laboratory tests, testing kits, and devices; however, these were not always available on hand. This study proposes a new approach in disease detection using machine learning algorithms by analyzing symptoms experienced by a person without requiring laboratory tests. Six supervised machine learning algorithms such as J48 decision tree, random forest, support vector machine, k-nearest neighbors, naïve Bayes algorithms, and artificial neural networks were applied in the “COVID-19 Symptoms and Presence Dataset” from Kaggle. Through hyperparameter optimization and 10-fold cross validation, we attained the highest possible performance of each algorithm. A comparative analysis was performed according to accuracy, sensitivity, specificity, and area under the ROC curve. Results show that random forest, support vector machine, k-nearest neighbors, and artificial neural networks outweighed other algorithms by attaining 98.84% accuracy, 100% sensitivity, 98.79% specificity, and 98.84% area under the ROC curve. Finally, we developed a web application that will allow users to select symptoms currently being experienced, and use it to predict the presence of COVID-19 through the developed prediction model. Based on this mechanism, the proposed method can effectively predict the presence or absence of COVID-19 in a person immediately without using laboratory tests, kits, and devices in a real-time manner.

In this study, we have established a feature extraction process for the classification of unknown genres of Turkish texts by using Turkish morphology. The proposed method considers the features as the word stems. The fact that the number of the features exceeds the practical computing limits each document represented by a number of features as in the document classes. Each word stem in a document analyzed in different classes and the sum of the usage frequency in the document classes given the feature value of that document. To speed up the process of extracting usage frequencies of word stems and analyzing it in different document classes Trie tree structure has been used. In this study, we have selected five different classes which are economy, healthy, magazine, sports and politics. The performance of the established method has been compared the bag of words approach by using naive Bayes, support vector machine, K-nearest neighbor, C 4.5 and random forest. The best performance achieved is 96.25% which has been observed using the naive Bayes with our new feature vectors.

Gesture recognition is a topic in computer science and language technology that aims to interpret human gestures with computer programs and many different algorithms. It can be seen as the way computers can understand human body language. Today, the main interaction tools between computers and humans are still the keyboard and mouse. Gesture recognition can be used as a tool for communication with the machine and interaction without any mechanical device such as a keyboard or mouse. In this paper, we present the results of a comparison of eight different machine learning (ML) classifiers in the task of human hand gesture recognition and classification to explore how to efficiently implement one or more tested ML algorithms on an 8-bit AVR microcontroller for on-line human gesture recognition with the intention to gesturally control the mobile robot. The 8-bit AVR microcontrollers are still widely used in the industry, but due to their lack of computational power and limited memory, it is a challenging task to efficiently implement ML algorithms on them for on-line classification. Gestures were recorded by using inertial sensors, gyroscopes, and accelerometers placed at the wrist and index finger. One thousand and eight hundred (1800) hand gestures were recorded and labelled. Six important features were defined for the identification of nine different hand gestures using eight different machine learning classifiers: Decision Tree (DT), Random Forests (RF), Logistic Regression (LR), Linear Discriminant Analysis (LDA), Support Vector Machine (SVM) with linear kernel, Naïve Bayes classifier (NB), K-Nearest Neighbours (KNN), and Stochastic Gradient Descent (SGD). All tested algorithms were ranged according to Precision, Recall, and F1-score (abb.: P-R-F1). The best algorithms were SVM (P-R-F1: 0.9865, 0.9861, and 0.0863), and RF (P-R-F1: 0.9863, 0.9861, and 0.0862), but their main disadvantage is their unusability for on-line implementations in 8-bit AVR microcontrollers, as proven in the paper. The next best algorithms have had only slightly poorer performance than SVM and RF: KNN (P-R-F1: 0.9835, 0.9833, and 0.9834) and LR (P-R-F1: 0.9810, 0.9810, and 0.9810). Regarding the implementation on 8-bit microcontrollers, KNN has proven to be inadequate, like SVM and RF. However, the analysis for LR has proved that this classifier could be efficiently implemented on targeted microcontrollers. Having in mind its high F1-score (comparable to SVM, RF, and KNN), this leads to the conclusion that the LR is the most suitable classifier among tested for on-line applications in resource-constrained environments, such as embedded devices based on 8-bit AVR microcontrollers, due to its lower computational complexity in comparison with other tested algorithms.

This paper introduces a specific approach for leaf classification based on machine learning (ML), transfer learning (TL), and convolutional neural network (CNN). The proposed method involves three stages, namely data preprocessing, feature extraction, and classification. This research uses images of leaves to distinguish between different plant species as leaves of one species differ from other species. Leaves are specific to each other via traits which include form, shade, texture, and margin. The dataset used for this experiment is the Swedish leaf dataset, a database of 15 different plant species with 1125 leaf images. Experimental results showed that random forest (RF) achieved a classification accuracy of 98.83% against other ML algorithms with a combination of grayscale images, HSV color moments, Hu moments, and Haralick features. The ResNet50 model gave us the best accuracy of 99.85% compared to other models. A CNN convolves learned features with input data and uses 2D convolutional layers. It means that this type of network is ideal for 2D image processing. We have built our own CNN model from scratch and managed to reach an accuracy of 98.04%.KeywordsMachine learningDeep learningTransfer learningLeaf classificationAccuracy

Support vector machines (SVMs) are supervised learning models that analyze data for classification or regression. For classification, SVM is widely used by selecting an optimal hyperplane that separates two classes. SVM has very good accuracy and extremally robust comparing with some other classification methods such as logistics linear regression, random forest, k-nearest neighbor and naïve model. However, working with large datasets can cause many problems such as time-consuming and inefficient results. In this paper, the SVM has been modified by using a stochastic Gradient descent process. The modified method, stochastic gradient descent SVM (SGD-SVM), checked by using two simulation datasets. Since the classification of different cancer types is important for cancer diagnosis and drug discovery, SGD-SVM is applied for classifying the most common leukemia cancer type dataset. The results that are gotten using SGD-SVM are much accurate than other results of many studies that used the same leukemia datasets.

Emotion plays a very significant role in human's everyday communication. Emotion recognition is an essential activity for a computer to understand the human inner state in the Internet of things (IoT)-based brain–computer interface (BCI) framework. Researchers nowadays are more intent toward emotion detection from electroencephalogram (EEG) signals. In recent years, many researchers have conducted various methods of emotion detection from physiological (EEG) signals. A variety of features are extracted from the EEG signals, and various classification methods are applied to these features. In this chapter, different classifiers based on machine learning such as K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and Gaussian Naive Bayes (GNB) with Principal Component Analysis (PCA) are used to simulate the relationship between emotional states and EEG signals of the brain. DEAP dataset is used to test the efficacy of these machine learning techniques for the analysis of emotions using physiological EEG signals and it is realized that SVM with PCA significantly improves accuracy in cases of valence, arousal, and dominance. In the case of liking, KNN with PCA gives better results in terms of accuracy. It is also found that SVM with PCA provides better classification accuracy i.e. 64.06% for valence, 64.30% for arousal, 69.50% for dominance, and 64.06% for liking, respectively.