13th Annual Conference of the International Speech Communication Association

Portland, OR, USA
September 9-13, 2012

Evaluation of a Sparse Representation-Based Classifier For Bird Phrase Classification Under Limited Data Conditions

Lee Ngee Tan (1), Kantapon Kaewtip (1), Martin L. Cody (2), Charles E. Taylor (2), Abeer Alwan (1)

(1) Department of Electrical Engineering; (2) Department of Ecology and Evolutionary Biology;
University of California, Los Angeles, CA, USA

This paper evaluates the performance of a sparse representation-based (SR) classifier for a limited data, bird phrase classification task. The evaluation database contains 32 unique phrases segmented from songs of the Cassinfs Vireo (Vireo cassinii). Spectrographic features were extracted from each phrase-segmented audio file, followed by dimension reduction using principal component analysis (PCA). A performance comparison to the nearest subspace (NS) and support vector machine (SVM) classifiers was conducted. The SR classifier outperforms the NS and SVM classifiers, with a maximum absolute improvement of 3.4% observed when there are only four tokens per phrase in the training set.

Index Terms: bird phrase classification, limited data, sparse representation, L1 minimization.

Full Paper

Bibliographic reference.  Tan, Lee Ngee / Kaewtip, Kantapon / Cody, Martin L. / Taylor, Charles E. / Alwan, Abeer (2012): "Evaluation of a sparse representation-based classifier for bird phrase classification under limited data conditions", In INTERSPEECH-2012, 2522-2525.