Pitch-density-based features and an SVM binary tree approach for multi-class audio classification in broadcast news

Abstract

Audio classification is an essential task in multimedia content analysis, which is a prerequisite to a variety of tasks such as segmentation, indexing and retrieval. This paper describes our study on multi-class audio classification on broadcast news, a popular multimedia repository with rich audio types. Motivated by the tonal regulations of music, we propose two pitch-densitybased features, namely average pitch-density (APD) and relative tonal power density (RTPD). We use an SVM binary tree (SVM-BT) to hierarchically classify an audio clip into five classes: pure speech, music, environment sound, speech with music and speech with environment sound. Since SVM is a binary classifier, we use the SVM-BT architecture to realize coarse-to-fine multi-class classification with high accuracy and efficiency. Experiments show that the proposed one-dimensional APD and RTPD features are able to achieve comparable accuracy with popular high-dimensional features in speech/music discrimination, and the SVM-BT approach demonstrates superior performance in multi-class audio classification. With the help of the pitch-density-based features, we can achieve a high average accuracy of 94.2% in the five-class audio classification task.

Albayzín-2014 evaluation: audio segmentation and classification in broadcast news domains

“…Frame-based features have also been proposed for segmenting and classifying BN audio into broad classes. As an example, two pitchdensity-based features are proposed in [23], the authors use short-time energy (STE) in [1,24,25], and harmonic features are used in [26][27][28]. The frame-based features can be directly used in the classifier.…”

Audio segmentation-by-classification approach based on factor analysis in broadcast news domain

“…More precisely, these features with other extended sets of features have been proposed for segmenting and classifying BN audio into broad classes. Among others, two pitch-density-based features are proposed in [11], short-time energy (STE) is used in [12][13][14], and harmonic features are used in [15][16][17]. The previously mentioned features are short-term characteristics because they are extracted within short periods of time (between 10 and 30 ms), usually known in the literature as frame-based features.…”

Sentence boundary detection in chinese broadcast news using conditional random fields and prosodic features

“…These story boundaries are certainly sentence boundaries. Therefore, we use an SVM binary tree approach [13] to detect music regions and whether …”