Abstract—Acoustic monitoring is commonly used in forest studies since sound recordings can capture information related to animal activity, human presence, and environmental changes over time. However, traditional monitoring methods based on manual observation or remote sensing are limited in detecting short and irregular acoustic events and often require continuous human effort. This work presents a deep learning-based approach, termed CLASED (CNN–LSTM–Attention-based Sound Event Detection). The framework combines convolutional neural networks for extracting spectral features with long short-term memory networks to model temporal variations in audio signals. An attention mechanism is incorporated to improve the contribution of relevant time segments during the classification process. A dataset of 3,000 forest-related audio recordings was prepared and processed through resampling, augmentation, and feature normalization before training. The developed model classifies audio signals into three categories: animal calls, human activity, and natural environmental sounds. Experimental results show an overall classification accuracy of 79%, with better performance observed for animal vocalizations and consistent performance across the remaining classes. The study demonstrates that integrating temporal modeling with attention can support reliable automated analysis of forest sound data and reduce dependence on manual monitoring

Index Terms—acoustic signal analysis, convolutional neural networks, deep learning for audio recognition, environmental sound classification, forest acoustic analysis, long short-term memory networks

Cite: Maddhigalla Lakshumaiah and D. S. Rao, "Deep Learning-Based Multi-Class Forest Sound Classification Using a CNN–LSTM–Attention Framework," International Journal of Electrical and Electronic Engineering & Telecommunications, vol. 15, no. 4, pp. 304-311, 2026. doi: 10.18178/ijeetc.15.4.304-311

Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).