Table 1.

Machine learning-based studies in plant stress under the Identification, Classification, Quantification, and Prediction (ICQP) paradigm.

Artificial Intelligence Technique	Algorithms	(ICQP) Application	Datasets	Model Plant Reported	Stressor	Reference
Deep Learning (image)	Convolutional neural networks (CNN), AlexNet, GoogLeNet, and Inception V3	Identification	1200 images acquired by camera under stress and non-stress conditions	Maize (Zea mays), okra (Abelmoschus esculentus), and soybean (Glycine max)	Water stress	Chandel et al. (2020) [113]
Unsupervised Machine learning	Least squares discriminant analysis (PLS-DA) and least-squares support vector machine (LS-SVM)	Identification	Hyperspectral images of the canopy of tobacco plants	Tobacco	Heavy metal stress Hg	Yu et al. (2021) [114]
Deep Learning (image)	CNN	Identification	1426 images of rice diseases and pests from paddy fields	Rice	Biotic stress	Rahman et al. (2020) [115]
Unsupervised Machine learning (video imaging)	Hidden Markov models (HMMs)	Identification and classification	Chlorophyll fluorescence (ChlF) digital profiles from GrowTech Inc.	Phaseolus vulgaris L. (Snap bean)	Stressor “level” groups (low, medium, and high stressed) and three stressor “type” categories (drought, nutrient, and chemical stress)	Blumenthal et al. (2020) [116]
Deep Learning (image)	CNN	Identification and Quantification	1747 smartphones images of arabica coffee leaves.	Arabica coffee	Biotic stress; leaf miner, rust, brown leaf spot, and Cercospora leaf spot	Esgario et al. (2020) [117]
Supervised Machine Learning, Partial Least Square Regression, Principal Component Analysis, and combined models	K-nearest neighbors (KNN)	Identification and classification	Spectral signature of leaf samples obtained with a visible, near-infrared spectrometer	Rice	Salt stress	Das et al. (2020) [118]
Supervised Machine Learning	ReliefF, support vector machine (SVM), recursive feature elimination (RFE), and random forest (RF)	Identification and classification	Hyperspectral images from four wheat lines	Wheat	Salt stress	Moghimi et al. (2018) [119]
Deep Learning (image)	CNN	Identification and classification	1575 images (smartphones, compact cameras, DSLR	Different plant specimens	Biotic stress	Arnal Barbedo (2019) [120]
Deep Learning	RF, SVM, multilayer perceptron (MLP)	Identification and classification	Hyperspectral images	Bromus inermis	Drought stress	Dao et al. (2021) [121]
Supervised Machine Learning	SVM	Identification and classification	RGB leave images from the Kaggle database	Brinjal leaves	Biotic stress	Karthickmanoj et al. (2021) [122]
Deep Learning (image)	Deep convolutional neural network (DCNN)	Identification, classification, and quantification	Collection of images of stressed and healthy soybean leaflets in the field	Soybean [Glycine max (L.) Merr.]	Bacterial blight (Pseudomonas savastanoi pv. glycinea), bacterial pustule (Xanthomonas axonopodis pv. glycines), sudden death syndrome (Fusarium virguliforme), septoria brown spot (Septoria glycines), frogeye leaf spot (Cercospora sojina), iron deficiency chlorosis, potassium deficiency, and herbicide injury	Ghosal et al. (2018) [123]
Supervised Machine Learning	RF, SVM, KNN	Classification and prediction	Real time terahertz time-domain spectroscopic data (THz-TDS)	Basil, coriander, parsley, baby-leaf, coffee, pea-	Water Stress	Zahid et al. (2022) [124]
Supervised Machine Learning	RF, artificial neural networks (ANN), and	Classification	Multispectral images	Maize	Water stress	Niu et al. (2021) [125]
Supervised Machine Learning	Confident multiple-choice learning	Identification and prediction	Gene expression time-series datasets	Arabidopsis thaliana	Heat, cold, salt, and drought	Kang et al. (2018) [126]
Deep Learning (image)	CNN	Classification	Images of Sorghum plant shoot from the Donald Danforth Plant Science Center.	Sorghum plants	Nitrogen deficiency	Azimi et al. (2021) [127]
Supervised Machine Learning	Decision tree (DT), SVM, and Naïve Bayes (NB)	Classification	Metabolite and protein content	Arabidopsis thaliana	Metabolic stress	Fürtauer et al. (2018) [128]
Supervised Machine Learning	SVM	Classification	Biweekly RGB, stereo and hyperspectral spatio-temporal images	Sugar beet plants	Abiotic stress conditions (drought and nitrogen deficiency) and one biotic stressor (weed)	Khanna et al. (2019) [129]
Supervised Machine Learning	Hierarchical models	Classification	5916 RGB images (493 plots including Plant Introduction (PI) accessions in different time points)	Soybean (Glycine max (L.) Merr.)	Iron deficiency chlorosis	Naik et al. (2017) [130]
Supervised Machine Learning	ANN, CNN, optimum-path forest, KNN, and SVM	Classification	Electrical signal under cold, low light and osmotic stimuli.	Soybean plants	Cold, low light, and osmotic stimuli.	Pereira et al. (2018) [131]
Supervised Machine Learning	RF	Classification	Hyperspectral dataset acquired from the Indian Agricultural Research Institute (IARI)	Wheat	Water stress	Mondal et al. (2019) [132]
Deep Learning (image)	CNN, SVM	Classification	65,184 labeled images from Github resources	Soybean	Biotic (fungal and bacterial diseases) and abiotic (nutrient deficiency and chemical injury) stresses	Venal et al. (2019) [133]
Supervised Machine Learning	MLP and probabilistic neural network (PNN)	Classification	16 maize and 17 wheat genomic and phenotypic datasets with different trait-environment combinations	Maize and Wheat	Drought	González-Camacho et al. (2016) [134]
Supervised Machine Learning	Decision tree (DT), SVM, and NB	Prediction	miRNA concentration.	Arabidopsis thaliana plants	Drought, salinity, cold, and heat	Vakilian (2020) [135]
Supervised Machine Learning	Ridge regression, LASSO, elastic net, RF, reproducing kernel Hilbert space, Bayes A and Bayes B	Prediction	A set of 29,619 cured Single Nucleotide Polymorphisms, genotyped across a panel of 240 maize inbred lines	Maize	Drought stress	Shikha et al. (2017) [136]
Deep Learning	CNN	Prediction	Three maize and six wheat data sets.	Maize and wheat	Environmental stress	Montesinos-López et al. (2018) [137]
Supervised Machine Learning	Genomic random regression	Prediction	Complete genotypes, molecular markers, and phenotypic traits of stressed and control groups.	Wheat	Environmental stress	Ly et al. (2018) [138]