Abstract—To support biodiversity conservations, plant classification studies, particularly from images, are necessary. This study explores the use of the deep convolutional neural network as a feature extractor to a plant classification problem. An original dataset consisting of images of seedlings of the three most important berry trees belonging to the Philippine indigenous plants was used. The result shows that as the network layers are getting deeper, they are becoming better at extracting discriminative features, such that, irrespective of classifier used their prediction performance keeps on improving. When the different layers were individually visualized, the features extracted were far from random, uninterpretable patterns. Rather, they show relevant properties that are capable of sorting patterns progressively from low to higher level. Hence, for classification problems bounded with the limitation of data, time, and computational hardware, leveraging the representational power of the deep convolutional neural network is very useful.
 

Index Terms—feature extraction, deep convolutional neural network, deep learning, AlexNet, plant classification, SVM

Cite: Jolitte A. Villaruz, "Deep Convolutional Neural Network Feature Extraction for Berry Trees Classification," Journal of Advances in Information Technology, Vol. 12, No. 3, pp. 226-233, August 2021. doi: 10.12720/jait.12.3.226-233

Copyright © 2021 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.