Convolutional neural network classification of Coleoptera through keras and tensorflow

Abstract

An application was developed to address the time-consuming and subjective nature of physical examination and human judgment to classify the diverse families of Coleoptera more efficiently. Specifically focusing on the Coleoptera families in the Philippines, the application utilizes a Convolutional Neural Network (CNN)-based model as its classification architecture. To enhance the accuracy of the CNN model, researchers gathered images of Coleoptera from various families. The program achieves an impressive 94.44% accuracy rate and has demonstrated successful results in classifying Coleoptera families. This application offers a more streamlined and reliable approach to Coleoptera classification, benefiting researchers and enthusiasts.