<![CDATA[This study compares the performance of two deep learning architectures, Xception and MobileNet, for classifying skipjack and mackerel tuna, with a focus on accuracy and computational efficiency. MobileNet achieved an impressive accuracy of 97%, with precision, recall, and F1-score all at 97%, and demonstrated a faster prediction time of 0.06 seconds, making it well-suited for real-time applications. In contrast, Xception achieved an accuracy of 93%, with a precision of 94%, recall of 93%, and an F1-score of 93%. However, its prediction time was slower at 0.13 seconds, indicating a higher computational complexity. Although Xception delivered substantial accuracy, MobileNet outperformed it in terms of efficiency, suggesting that MobileNet is better suited for applications with limited resources or time constraints. The results indicated that MobileNet's lightweight architecture makes it ideal for mobile or embedded systems. At the same time, Xception's more complex structure may be advantageous for tasks that require higher precision in image processing. This research makes a significant contribution to the development of deep learning-based methods for fish species classification, offering improvements in both accuracy and speed.]]>
Copyrights © 2025
