Table 2 Model summary.
From: A lightweight deep learning method for medicinal leaf image classification using feature fusion
Refs. | Method | Dataset | Accuracy |
|---|---|---|---|
GoogLeNet + linear SVM |  | 87.34%. | |
Convolution neural network | Â | 86% | |
Five-layered convolutional neural network (CNN) | Flavia leaf dataset Swedish leaf dataset | 98.22%. | |
CNN-LSTM network through 20 layers | Â | 95.06%. | |
MobileNetV2 | Â | 98.97 | |
Dual-path CNN (DP-CNN) | Â | 95.67% | |
Dual-path CNN model | 14 plant speicies. | 77.1% | |
AlexNet, VGG-19, GoogLeNet, ResNet50, and MobileNetV2 | Leafsnap image dataset | 92.3% | |
5-Layer CNN architecture | Flavia leaf dataset | 95.5 98.2 | |
GoogleNet, VGGNet, and AlexNet | LIFECLEF 2015 dataset | 80% | |
Two AlexNets pretrained models | Â | 97.3% | |
ResNet152 and Inception-ResNetv2 architectures with LBP | Swedish leaf dataset | 97% | |
Seven-layer CNN | Flavia dataset | 94% | |
AlexNet and GoogLeNet | Swedish leaf dataset | 94% 98% 96% | |
17-Layer CNN architecture | Â | 97.9% | |
VGG19 architecture by a logistic regression classifier | Swedish leaf datasets | 96% 96% 97.85%, | |
AousethNet | Mendeley dataset (MD2020) | 98% |
