SOUND RECOGNITION OF SPECIES USING TENSOR FLOW

Abstract

Animal sound recognition is the technological advancement in the audio recognition using machine
Learning and deep learning. The audio recognition is traditionally focussed on the speech. The main purpose of animal
sound recognition is to recognise the emotions of the species as animals and birds are tending to change their activities
as well as their habitats due to the adverse effects on the environment or due to other natural or man-made calamities.
The best way to monitor the species is by audio recognition. We train the machines by pre-recorded audio files. And
these audio files should have no background noise or echoes to avoid overlapping.
We here use convolutional neural networks using tensor flow and train the audio datasets of the species. The
convolutional neural networks work on the frequency cepstral coefficients that are extracted from the audio datasets.
The manuscript serves as a technical paper showing how the model works to achieve the desired result.