Machine learning is advancing classifiers by differentiating complex sample spaces using significant stores of collected data. Classifying bird species by vocalization is an active area of research with ecological applications. Increasingly accurate identification of bird species can improve crowd-sourced data for ornithology research. Current technology achieves maximum 85.4% AUC for 88 classifications.
The avian classifier acquires data through automated gathering of Creative Commons licensed bird vocalization records. Training sets select highly rated recordings with few background species and each recording undergoes band pass and median threshold filters. Audio fingerprints are created through fast combinatorial hashing of relative positions of spectrogram peaks trading FFT window size and hashing tolerance. Captured hashes serve as truth data in the template matching artificial intelligence model. New recordings undergo the same filtering and fingerprinting prior to matching and scoring for classification. Small tests of 8 recordings of 4 species indicate distinct differences and 75% accurate classification by large margins in the scoring histograms. Large scale tests perform at 50% due to similarities of calls between species, inclusion of a significant number of noise points in the training data, and recordings chosen to exhibit the technique’s weaknesses including recording of a single note, high noise recordings, and cross species similar vocalizations.
|Commitee:||Aliasgari, Mehrdad, Penzenstadler, Birgit|
|School:||California State University, Long Beach|
|Department:||Computer Engineering and Computer Science|
|School Location:||United States -- California|
|Source:||MAI 56/04M(E), Masters Abstracts International|
|Keywords:||Avian, Birdcall, Classifier, Fingerprinting, Shazam, Templating|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be