Two convolutional neural networks for bird detection in audio signals

Two convolutional neural networks for bird detection in audio signals. Grill, T. & Schlüter, J. In 2017 25th European Signal Processing Conference (EUSIPCO), pages 1764-1768, Aug, 2017.

Paper doi abstract bibtex

We present and compare two approaches to detect the presence of bird calls in audio recordings using convolutional neural networks on mel spectrograms. In a signal processing challenge using environmental recordings from three very different sources, only two of them available for supervised training, we obtained an Area Under Curve (AUC) measure of 89% on the hidden test set, higher than any other contestant. By comparing multiple variations of our systems, we find that despite very different architectures, both approaches can be tuned to perform equally well. Further improvements will likely require a radically different approach to dealing with the discrepancy between data sources.

@InProceedings{8081512,
  author = {T. Grill and J. Schlüter},
  booktitle = {2017 25th European Signal Processing Conference (EUSIPCO)},
  title = {Two convolutional neural networks for bird detection in audio signals},
  year = {2017},
  pages = {1764-1768},
  abstract = {We present and compare two approaches to detect the presence of bird calls in audio recordings using convolutional neural networks on mel spectrograms. In a signal processing challenge using environmental recordings from three very different sources, only two of them available for supervised training, we obtained an Area Under Curve (AUC) measure of 89% on the hidden test set, higher than any other contestant. By comparing multiple variations of our systems, we find that despite very different architectures, both approaches can be tuned to perform equally well. Further improvements will likely require a radically different approach to dealing with the discrepancy between data sources.},
  keywords = {audio recording;audio signal processing;feature extraction;learning (artificial intelligence);neural nets;convolutional neural networks;bird detection;audio signals;bird calls;audio recordings;mel spectrograms;environmental recordings;supervised training;area under curve measure;AUC measure;Birds;Training;Spectrogram;Convolution;Training data;Computer architecture},
  doi = {10.23919/EUSIPCO.2017.8081512},
  issn = {2076-1465},
  month = {Aug},
  url = {https://www.eurasip.org/proceedings/eusipco/eusipco2017/papers/1570347092.pdf},
}

Downloads: 0

{"_id":"4ZfwSpitWfqurMxTj","bibbaseid":"grill-schlter-twoconvolutionalneuralnetworksforbirddetectioninaudiosignals-2017","authorIDs":[],"author_short":["Grill, T.","Schlüter, J."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["T."],"propositions":[],"lastnames":["Grill"],"suffixes":[]},{"firstnames":["J."],"propositions":[],"lastnames":["Schlüter"],"suffixes":[]}],"booktitle":"2017 25th European Signal Processing Conference (EUSIPCO)","title":"Two convolutional neural networks for bird detection in audio signals","year":"2017","pages":"1764-1768","abstract":"We present and compare two approaches to detect the presence of bird calls in audio recordings using convolutional neural networks on mel spectrograms. In a signal processing challenge using environmental recordings from three very different sources, only two of them available for supervised training, we obtained an Area Under Curve (AUC) measure of 89% on the hidden test set, higher than any other contestant. By comparing multiple variations of our systems, we find that despite very different architectures, both approaches can be tuned to perform equally well. Further improvements will likely require a radically different approach to dealing with the discrepancy between data sources.","keywords":"audio recording;audio signal processing;feature extraction;learning (artificial intelligence);neural nets;convolutional neural networks;bird detection;audio signals;bird calls;audio recordings;mel spectrograms;environmental recordings;supervised training;area under curve measure;AUC measure;Birds;Training;Spectrogram;Convolution;Training data;Computer architecture","doi":"10.23919/EUSIPCO.2017.8081512","issn":"2076-1465","month":"Aug","url":"https://www.eurasip.org/proceedings/eusipco/eusipco2017/papers/1570347092.pdf","bibtex":"@InProceedings{8081512,\n author = {T. Grill and J. Schlüter},\n booktitle = {2017 25th European Signal Processing Conference (EUSIPCO)},\n title = {Two convolutional neural networks for bird detection in audio signals},\n year = {2017},\n pages = {1764-1768},\n abstract = {We present and compare two approaches to detect the presence of bird calls in audio recordings using convolutional neural networks on mel spectrograms. In a signal processing challenge using environmental recordings from three very different sources, only two of them available for supervised training, we obtained an Area Under Curve (AUC) measure of 89% on the hidden test set, higher than any other contestant. By comparing multiple variations of our systems, we find that despite very different architectures, both approaches can be tuned to perform equally well. Further improvements will likely require a radically different approach to dealing with the discrepancy between data sources.},\n keywords = {audio recording;audio signal processing;feature extraction;learning (artificial intelligence);neural nets;convolutional neural networks;bird detection;audio signals;bird calls;audio recordings;mel spectrograms;environmental recordings;supervised training;area under curve measure;AUC measure;Birds;Training;Spectrogram;Convolution;Training data;Computer architecture},\n doi = {10.23919/EUSIPCO.2017.8081512},\n issn = {2076-1465},\n month = {Aug},\n url = {https://www.eurasip.org/proceedings/eusipco/eusipco2017/papers/1570347092.pdf},\n}\n\n","author_short":["Grill, T.","Schlüter, J."],"key":"8081512","id":"8081512","bibbaseid":"grill-schlter-twoconvolutionalneuralnetworksforbirddetectioninaudiosignals-2017","role":"author","urls":{"Paper":"https://www.eurasip.org/proceedings/eusipco/eusipco2017/papers/1570347092.pdf"},"keyword":["audio recording;audio signal processing;feature extraction;learning (artificial intelligence);neural nets;convolutional neural networks;bird detection;audio signals;bird calls;audio recordings;mel spectrograms;environmental recordings;supervised training;area under curve measure;AUC measure;Birds;Training;Spectrogram;Convolution;Training data;Computer architecture"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/Roznn/EUSIPCO/main/eusipco2017url.bib","creationDate":"2021-02-13T16:38:25.699Z","downloads":0,"keywords":["audio recording;audio signal processing;feature extraction;learning (artificial intelligence);neural nets;convolutional neural networks;bird detection;audio signals;bird calls;audio recordings;mel spectrograms;environmental recordings;supervised training;area under curve measure;auc measure;birds;training;spectrogram;convolution;training data;computer architecture"],"search_terms":["two","convolutional","neural","networks","bird","detection","audio","signals","grill","schlüter"],"title":"Two convolutional neural networks for bird detection in audio signals","year":2017,"dataSources":["2MNbFYjMYTD6z7ExY","uP2aT6Qs8sfZJ6s8b"]}