Audio Event Recognition Involving Animals and Bird Species Using Machine Learning

The ecosystem is a home for a variety of living things that coexist perfectly. The ecology is stable when two species coexist in their habitats, but as a result of globalization and industrialization, many creatures are losing their natural habitats and moving into human homes in quest of safety. According to the Wildlife Protection Society of India, poaching caused a 34% increase in the number of leopard andtiger deaths in 2019. Advancements in machine learning for voice recognition are improving, making it crucial to monitor and preserve living organisms like animals and birds. Environmental changes and natural disasters change their habitats, making continuous monitoring essential. Auditory recognition is the most effective method for monitoring endangered species, as it can be challenging for humans to classify sound. Audio recognition is possible through strong signals and analysis procedures, using pre-recorded audio or live recordings. Machines can eliminate echoes and background noise, focusing on syllables as the smallest sound in machine learning. ML techniques train computers using labeled datasets of interest species. Mel-frequency cepstral coefficients are derived from audio files, and audio event detection (AED) is analyzed using the Hidden Markov Model in a machine learning algorithm demonstrating the system's development and providing desired results.