Hi José,

BirdNET is one of the most common ones! The model is open, or there are also no-code interfaces that you can use as well.

Perch is another one (from Google hDeepMind) - good article about it too here.

The OpenSoundscape python package is also great for training your own CNNs.

Hi Jose,

The links that Carly posted are great resources. Just adding a couple other no-coding-required options to this discussion:

PAMGuard is good for marine bioacoustic classification. It has classifiers already for many species that produce whistles and clicks. It can also be used to create custom classifiers. The program can have a steep learning curve so keep that in mind.

Arbimon is a cloud-based tool for data storage and bioacoustic classification. It has some existing classification models and can be used to create custom ones.

BirdNET Analyzer uses the BirdNET model with an easy to use GUI. Can be used as is or allows for custom classifiers using transfer learning from original BirdNET embeddings.

Cheers,

Jesse

Hi Jose,

You might be interested in BriteKit, which helps you build bioacoustic recognizers using deep learning.