Fauzi, Fauzi and Permanasari, Adhistya Erna and Akhmad Setiawan, Noor (2021) Butterfly Image Classification Using Convolutional Neural Network (CNN). In: International Conference on Electronics Representation and Algorithm (ICERA).
Butterfly_Image_Classification_Using_Convolutional_Neural_Network_CNN.pdf
Restricted to Registered users only
Download (1MB) | Request a copy
Abstract
Butterflies have an aesthetic, ecosystem, educational, health, economic, and scientific values which make them essential to be researched. They are, however, difficult to recognize due to several varieties and patterns and this means there is a need to group based on type for easier recognition. Moreover, one of the machine learning-based methods currently being used in the image classification process is the Convolutional Neural Network (CNN) which has been discovered to have the ability to produce accurate results. This study, therefore, proposes a CNN-based model built from scratch to classify the butterfly images. This was initiated with a simple 4-Conv model followed by a second experiment in the form of a 7-Conv model. Meanwhile, the models were validated using the split validation technique which divides them into three parts including the train, validation, and test, and later evaluated through a confusion matrix. The experimental results showed the proposed method obtained an average accuracy of 0.9744 with the highest achievement recorded at 0.9865 accuracy, 0.0490 loss, 0.9797 validation accuracy, 0.1386 validation loss, 78 minutes, and 0.98 average evaluation. The evaluation of ten types of butterflies using confusion matrix also showed the proposed 7-Conv model successfully predicted the image accurately with an average accuracy of 0.94. © 2021 IEEE.
Item Type: | Conference or Workshop Item (Paper) |
---|---|
Additional Information: | Cited by: 8 |
Uncontrolled Keywords: | Convolution; Image classification; Image recognition; Image resolution; Machine learning; Aesthetic value; Butterfly identification; Confusion matrix; Convolutional neural network; Economic values; Health economics; Images classification; Images processing; Machine-learning; Convolutional neural networks |
Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering |
Divisions: | Faculty of Engineering > Electronics Engineering Department |
Depositing User: | Sri JUNANDI |
Date Deposited: | 29 Oct 2024 06:15 |
Last Modified: | 29 Oct 2024 06:15 |
URI: | https://ir.lib.ugm.ac.id/id/eprint/8417 |