Year: 2024
Type: article
Abstract: Seashell classification presents significant challenges in image processing, particularly in distinguishing between blood shells (Anadara granosa) and feather mussels (Anadara antiquata). This study leverages deep learning and computer vision techniques to develop a classification model for seashell images using Convolutional Neural Networks (CNN). Additionally, we propose the RunCNN method to compare its performance with CNN. The research involves collecting a large dataset of blood shells and feather mussels, preprocessing the data, training the models, and evaluating their performance. Experimental results demonstrate that the CNN-based model achieves 87% accuracy, while the RunCNN method achieves 82% accuracy. Both models exhibit low loss, indicating their effectiveness in classifying seashell images. These findings highlight the potential of deep learning approaches for accurate and efficient seashell classification, with CNN outperforming RunCNN in this context.
Authors: Rike Pradila, Yuliana Aprillia Sahuburua
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4411447782