I have tons of audio recordings, and I wish to detect specific bird species within those files. I am currently evaluating the possibilities, and wish to have the opinion on what are the best systems for the task in R.
I would say any standard classifier (KNN, Decision trees and SVM) can do the job as long as the right features are extracted. Depending on the size of the recording database I would recommend Artificial Neural Network (ANN) classifier (not available in R) for larger databases and KNN for smaller ones.
If you don't have to use R, check this out:
http://dcase.community/challenge2019/index
I seem to remember a newspaper article (unfortunately in Finnish) from a few years back that talked about an app that was quite good at doing bird species recognition. Might have been based on this work http://legacy.spa.aalto.fi/research/avesound/pubs/selin_mst.pdf
And in parallel to what Shahab said, would be worth using something other than R for the task, if that's at all possible.
Cheers, Pertti
You could try to apply template matching using the package monitoR.
Katz, J., Hafner, S. D., & Donovan, T. (2016). Tools for automated acoustic monitoring within the R package monitoR. Bioacoustics, 25(2), 197-210.
Hope this help!
Cheers
Simon
Thank you all for the answers and suggestions. I have tried some of the available packages, and monitoR() was quite a nice surprise. I took me some time to get it properly, specially to apply it to a large data set, but it works quite well.
Did you already see Seppo Fagerlunds thesis? https://aalto.finna.fi/Record/aaltodoc.123456789_14415
- Badges
- Science topic
- Similar topics
- Filing