SVM and KNN Based CNN Architectures for Plant Classification

Automatic plant classification through leaf is a classical problem in Computer Vision. Plants challenging due to the introduction of new species with similar pattern and look-a-like. Many efforts are made automate using leaf, flower, bark, or stem. After much effort, it has been proven that most reliable source for classification. But identify help structure because shows similarity morphological variations, like sizes, textures, shapes, venation. Therefore, required normalize all leaves into same size get better performance. Convolutional Neural Networks (CNN) provides fair amount accuracy when classified this approach. performance can be improved by classifying traditional approach after applying CNN. In paper, two approaches, namely CNN + Support Vector Machine (SVM) K-Nearest Neighbors (kNN) used on 3 datasets, LeafSnap dataset, Flavia Dataset, MalayaKew Dataset. The datasets augmented take care possibilities. assessments correlations predetermined feature extractor models given. kNN managed reach maximum 99.5%, 97.4%, 80.04%, respectively, three datasets.