Music Annotation and Retrieval using Unlabeled Exemplars: Correlation and Sparse Codes

Abstract

Tagging music signals with semantic labels such as genres, moods and instruments is important for content-based music retrieval and recommendation. While considerable effort has been made, automatic music annotation is still considered challenging due to the difficulty of extracting good audio features that capture the characteristics of different tags. To address this issue, we present in this letter two exemplar-based approaches that represent the content of a music clip by referring to a large set of unlabeled audio exemplars. The first approach represents a music clip by the set of audio exemplars that is highly correlated with the short-time feature vectors of the clip, whereas the second approach represents a music clip as sparse linear combinations of its short-time feature vectors over the audio exemplars. Music annotation is then performed by learning the relevance of the audio examples to different tags using labeled data. These two approaches effectively capitalize the availability of unlabeled data to explore the commonality of music signals to find out tag-specific acoustic patterns, without domain knowledge and feature design. Evaluation on the CAL10k music genre tagging dataset for tag-based music retrieval shows that, with thousands of unlabeled audio examples randomly drawn from the Million Song Dataset, the proposed approaches lead to remarkably higher precision rates than existing approaches.

[Retracted] Research on the Sustainability of Traditional Music and the Adaptability of Ecological Environment

“…The music sustainable storage technology can be used to realize the in-depth analysis of the stored traditional music data, provide users with valuable information data, and then complete cultural exchanges and applications around the world but how to obtain specific content from the initial humanistic data lacking the definition of traditional music content has become a huge challenge for the adaptive evaluation method of the current traditional music [ 4 ]. Traditional music is inevitably influenced by today's economic laws and is subjected to the competition test of survival of the fittest [ 5 – 8 ]. Those folk and national traditional music projects that can serve commercial activities or have a certain commercial development value are constantly excavated.…”

BMNet-5: A Novel Approach of Neural Network to Classify the Genre of Bengali Music Based on Audio Features

“…The correlation coefficient is used to determine the linear dependence between the characteristics of two signals. [21]. Keeping the correlation in consideration they extracted timbral feature, rhythmic feature, and time-frequency features and used different parameters to evaluate their proposed models.…”

Optimal classification of music genres based on acoustic and visual features

“…Evaluate 𝛽 t (f) as in Equation (19). The compressing factor 𝛽 t (f) is modeled as: initially, the smooth spectra K t (f) are modeled through low pass filtering K t (f) in LF(f), and t subsequently compressing factor 𝛽 t (f) is modeled.…”

“…16,17 The motivation of this research is to find a way to overcome some of the flaws in conventional approaches. [18][19][20][21][22] The computational cost of the suggested method is lower, and multimodal information fusion is addressed.…”

“…Nevertheless, analyses on diverse superior DNN schemes are relatively few, it is complicated to substantiate which technique is the finest one for classifying music genres 16,17 . The motivation of this research is to find a way to overcome some of the flaws in conventional approaches 18‐22 . The computational cost of the suggested method is lower, and multimodal information fusion is addressed.…”