Detection and multi-class classification of freshwater fish species using space-to-depth convolution and YOLOv11

Key Points

• Detection accuracy improves with depth convolution techniques, enhancing identification efforts across habitats.
• Fifty-six species were classified correctly with an average accuracy of 85%, showcasing the model's potential.
• Analysis deployed a YOLOv11 deep learning framework for real-time classification of fish species in freshwater environments.
• Findings highlight the need for scalable identification solutions, indicating future applications in biodiversity monitoring.