Table 1 Detection with machine learning method.
From: A lightweight deep learning method for medicinal leaf image classification using feature fusion
Refs. | Method | Dataset | Accuracy | Summary |
|---|---|---|---|---|
SVM classifier | Chinese medicinal leaf | 93.3% | Custom feature extraction with a low detection rate. | |
Randon forest | 30 images of 24 plant leaf | 90.1% | Low detection rate due less amount of data avaiable | |
Logistic regression, KNN, linear discriminant analysis, classification and regression trees, SVM, and NN | Philippine herbal medicine plants | 98.6% | Leaf shape and venation structure features were utilized to identify medicinal leaf | |
Neural networks | Dataset containing 50 medicinal | 93.3% | Three handcrafted feature such as Texture, colour, and shape were utilized which minimize accuracy | |
SVM | Ayurvedic medicinal plant | 96.66% | Morphological features | |
Extreme learning machine (ELM) with KNN, DT, SVM, NB classifier, | Fisher’s iris plant, | 97% and 96% | Histogram for feature representation | |
SVM, KNN, NB, MLP, RF and BT algorithms | 25 herbal leaf, fruit, and vegetable species | 85.82 | Shape, texture, and colour | |
Multilayer perceptron, random forest, basic logistic, logit-boost, and bagging | Six forms of medicinal plant leaves | 99.01% | chi-square feature selection strategy | |
ANN model | 20 diverse Chinese medicinal plants | 98.3% | Â | |
Multiclass-SVM | Swedish leaf dataset | 93.26%. | Texture and colour features |
