2026-05-04T19:55:10.215Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:3:11664 2025-10-28T15:03:23Z 3

Convolutional Network Optimization for Call Identification of African Bird Species Using Modified Metaheuristics, Chapter in AIS Algorithms for Intelligent Systems: PCCDA 2025: International Conference on Paradigms of Communication, Computing and Data Analytics, Springer 2025 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/3/11664 https://link.springer.com/chapter/10.1007/978-981-96-6843-4_1 S. Malisic M. Protic V. Markovic S. Tedic L. Jovanovic N. Bacanin M. Zivkovic Proper bird calls classification is crucial for biodiversity monitoring, as it provides a way to evaluate health of ecosystem health, allowing species distribution tracking and detecting changes in populations, which is particularly important for species that are hard to monitor visually. Although difficult to being hard to be located visually, these species often have distinctive vocalizations, enabling detection of their presence by identifying their calls. Consequently, long-term audio data collection of audio recordings may help in the early detection of population changes, which indicates shifts in the health of a particular ecosystem. This study explores the potential of convolutional neural networks (CNNs) optimized by metaheuristics algorithms to accurately identify the calls of three distinctive African bird species, sharing a common habitat. As part of this research, an altered variant of the renowned reptile search algorithm is proposed to address the known issues of the baseline algorithm known issues, and is applied to tune the CNN for this specific task. The proposed methodology achieved highly promising outcomes, with the highest achieved classification accuracy of 85.71. bookPart Springer, Singapore 1 15 10.1007/978-981-96-6843-4_1 AIS Algorithms for Intelligent Systems: PCCDA 2025: Proceedings of International Conference on Paradigms of Communication, Computing and Data Analytics