Advanced ML Project : An Orca Call classifier using mel-spectrograms as audio representations to detect Killer whales
I have used the Killer Whale (Orca) Call Audios for making the dataset from the FOSS Organisation OrcaSound, which has hydrophones in Seattle to listen to the Southern Resident Killer Whales. The audio has been downloaded from AWS Bucket and preprocessed to obtain spectrograms.
Dataset Size I have used is about 240 Audio Clips. The Instructions to download Audio Clips is given in the page below :
(https://github.com/orcasound/orcadata/wiki/Pod.Cast-data-archive)
.png)