3,551 Hits in 7.6 sec

Evaluation Of Album Effect For Feature Selection In Music Genre Recognition

Igor Vatolkin, Günter Rudolph, Claus Weihs
2015 Zenodo  
Such effect was observed in [30] for the recognition of genres. Also the tags of songs belonging to the same albums may have higher co-occurrences as inspected in [20] .  ...  In music classification, multi-objective evolutionary feature selection was introduced in [40] for genre categorisation and later for the recognition of instruments [41] .  ... 
doi:10.5281/zenodo.1416327 fatcat:6nmlwn5pxza3bd4vtiubi5j2f4

Feature Selection Pitfalls And Music Classification

Rebecca Fiebrink, Ichiro Fujinaga
2006 Zenodo  
Acknowledgments We gratefully acknowledge support from the McGill University Max Stern Fellowship in Music and the Canada Foundation for Innovation.  ...  That work used the evaluation methodology of [7] to show that feature selection was effective in improving classification accuracy on both musical and non-musical data.  ...  We wondered whether feature selection would still offer any benefit for other problems in music involving many features, such as audio genre classification.  ... 
doi:10.5281/zenodo.1415143 fatcat:nbq6ohxdpvfwxai7vmctkc4sey

FMA: A Dataset For Music Analysis [article]

Michaël Defferrard, Kirell Benzi, Pierre Vandergheynst, Xavier Bresson
2017 arXiv   pre-print
We introduce the Free Music Archive (FMA), an open and easily accessible dataset suitable for evaluating several tasks in MIR, a field concerned with browsing, searching, and organizing large music collections  ...  We here describe the dataset and how it was created, propose a train/validation/test split and three subsets, discuss some suitable MIR tasks, and evaluate some baselines for genre recognition.  ...  An artist filter for artists to be part of one set only, thus avoiding any artist and album effect.  ... 
arXiv:1612.01840v3 fatcat:i7hmi4pp2rbsvii66xya3px3hm

Speech-Controlled Media File Selection on Embedded Systems

Yu-Fang H. Wang, Stefan W. Hamerich, Marcus E. Hennecke, Volker Schubert
2005 SIGDIAL Conferences  
In addition to basic commands such as "next" or "repeat" one main feature of the system is the selection of titles, artists, albums, genres, or composers by speech.  ...  We will describe the implemented dialog and discuss challenges for a real-world application. The findings and considerations of the paper easily extend to general audio media.  ...  For example it should be possible to play songs from a certain artist or all titles of a particular album or music of a certain genre.  ... 
dblp:conf/sigdial/WangHHS05 fatcat:6xgpxoy3k5azpib4aiopznlbfy

Fma: A Dataset For Music Analysis

Michaël Defferrard, Kirell Benzi, Pierre Vandergheynst, Xavier Bresson
2017 Zenodo  
An artist filter for artists to be part of one set only, thus avoiding any artist and album effect.  ...  various features and classifiers for top genre recognition on the medium subset.  ...  SWITCH and EPFL for hosting the dataset within the context of the SCALE-UP project, funded in part by the swissuniversities SUC P-2 program. Xavier Bresson is supported by NRF Fellowship NRFF2017-10.  ... 
doi:10.5281/zenodo.1414727 fatcat:c4d7sdijznfctkhcoesvnkqhr4

Improving Interpretable Genre Recognizing with Audio Future Statistics Based on Zygonic Theory

Igor Vatolkin
2021 Zenodo  
Automatic music genre recognition helps to organize music collections and discover new music pieces.  ...  In this work, we propose novel and interpretable statistics derived from zygonic theory and estimated for discretized audio features.  ...  Acknowledgments This work was funded by the German Research Foundation (DFG), project 336599081 "Evolutionary optimisation for interpretable music segmentation and music categorisation based on discretised  ... 
doi:10.5281/zenodo.5724304 fatcat:jslzoswnzvdmliwi2duzrpw7de

Multi-Objective Investigation of Six Feature Source Types for Multi-Modal Music Classification

Igor Vatolkin, Cory McKay
2022 Transactions of the International Society for Music Information Retrieval  
With the intent of measuring the individual impact of different feature groups on different categories, we propose two evaluation criteria based on "non-dominated hypervolumes": multi-group feature "importance  ...  Our results highlight the potential of combining features extracted from different modalities, and can provide insight on the relative significance of different modalities and features in different contexts  ...  The experiments were carried out on the Linux HPC cluster at TU Dortmund (LiDO3), partially funded in the course of the Large-Scale Equipment Initiative by the German Research Foundation (DFG) as grant  ... 
doi:10.5334/tismir.67 fatcat:glhczbyvfjgabmergrixspzs3m

A Picture is Worth a Thousand Songs

Alexander Schindler
2014 Proceedings of the 1st International Workshop on Digital Libraries for Musicology - DLfM '14  
Yet, music is often searched for a specific intention such as music for workout, to focus or for a comfortable dinner with friends.  ...  Yet, extensive research in all disciplinary fields of music research, such as music psychology, musicology and information technologies, is required to identify correlations between the acoustic and the  ...  The following tasks are prevalent in the Music Information Retrieval Evaluation eXchange (MIREX) [20] : Genre Recognition: Automatic genre recognition has been extensively studied in the MIR domain [  ... 
doi:10.1145/2660168.2660185 dblp:conf/jcdl/Schindler14 fatcat:5hoblranurhy5hhku6klcq3ksy

Multi-Modal Music Information Retrieval: Augmenting Audio-Analysis with Visual Computing for Improved Music Video Analysis [article]

Alexander Schindler
2020 arXiv   pre-print
In a further consequence it can be concluded that this visual information is music related and thus should be beneficial for the corresponding MIR tasks such as music genre classification or mood recognition  ...  Additionally, new visual features are introduced capturing rhythmic visual patterns. In all of these experiments the audio-based results serve as benchmark for the visual and audio-visual approaches.  ...  Chapter 8 provides an in-depth evaluation of primary, low-level and affective features towards their effectiveness in music genre classification tasks.  ... 
arXiv:2002.00251v1 fatcat:6cz6rivc3fbg7fahdsnokxfrk4

You Can Judge an Artist by an Album Cover: Using Images for Music Annotation

Janis Libeks, Douglas Turnbull
2011 IEEE Multimedia  
Such visual information may be helpful for improving music discovery in terms of the quality of recommendations, the efficiency of the search process, and the aesthetics of the multimedia experience.  ...  While the perception of music tends to focus on our acoustic listening experience, the image of an artist can play a role in how we categorize (and thus judge) the artistic work.  ...  More details on these standard IR metrics can be found in Chapter 8 of [13] . To evaluate image similarity features, we compare the averages of the AUC and AP over all 50 genre tags.  ... 
doi:10.1109/mmul.2011.1 fatcat:sjpr4alztjgzrjuu3udhew337a

Evaluation of state of the art for genre classification in large datasets

Vibhor Bajpai, Dmitry Bogdanov, Alastair Porter
2018 Zenodo  
The goal of this thesis is to evaluate state of the art methods for genre classification on some popular genre datasets and provide an alternate dataset for the music community to use for the task of genre  ...  This thesis is an attempt to create models with better feature selection, and creating a dataset leveraging the publicly available Jamendo audio set for the task of genre classification.  ...  Out of all the available audio content in the Jamendo music collection, a total number of 5544 songs were selected representing 8 genres.  ... 
doi:10.5281/zenodo.3770036 fatcat:ratzdk37nvaz7podnw37kayuei

Multi-Task Music Representation Learning from Multi-Label Embeddings [article]

Alexander Schindler, Peter Knees
2019 arXiv   pre-print
The approach is evaluated in a multi-task scenario for which we introduce four large multi-tag annotations for the Million Song Dataset for the music properties genres, styles, moods, and themes.  ...  This paper presents a novel approach to music representation learning. Triplet loss based networks have become popular for representation learning in various multimedia retrieval domains.  ...  Acknowledgements This article has been made possible partly by received funding from the European Unions Horizon 2020 research and innovation program in the context of the VICTORIA project under grant  ... 
arXiv:1909.07730v1 fatcat:j74dxcgk7ranvifwagczcrxtuq

Music Artist Classification with Convolutional Recurrent Neural Networks [article]

Zain Nasrullah, Yue Zhao
2019 arXiv   pre-print
Previous attempts at music artist classification use frame level audio features which summarize frequency content within short intervals of time.  ...  Comparatively, more recent music information retrieval tasks take advantage of temporal structure in audio spectrograms using deep convolutional and recurrent models.  ...  INTRODUCTION Music information retrieval (MIR) encompasses most audio analysis tasks such as genre classification, song identification, chord recognition, sound event detection, mood detection and feature  ... 
arXiv:1901.04555v2 fatcat:gwcgvg6qvjaphm4xe2sb4qnvl4

Evaluation of Jamendo Database as Training Set for Automatic Genre Recognition [chapter]

Mariusz Kleć
2012 Lecture Notes in Computer Science  
The approach described in this paper is to use the public Jamendo database for training the chosen classifier for genre recognition task.  ...  This paper focuses on music classification by genre which is a type of supervised learning oriented problem.  ...  I would like to thank Danijel Korzinek for his help and good advice and ideas. I am also very grateful to Alicja Wieczorkowska and Krzysztof Marasek for their support. References  ... 
doi:10.1007/978-3-642-35455-7_27 fatcat:ge7jjidpindmvnyur7fqfsw3ei

Efficient multivariate sequence classification [article]

Pavel P. Kuksa
2014 arXiv   pre-print
In this work, we consider the problem of the multivariate sequence classification such as classification of multivariate music sequences, or multidimensional protein sequence representations.  ...  Typical kernel functions for sequences in these domains (e.g., bag-of-words, mismatch, or subsequence kernels) are restricted to discrete univariate (i.e. one-dimensional) string data, such as sequences  ...  For music artist recognition, we follow a 6-fold leave-one-album-out validation procedure proposed in the previous work [7] .  ... 
arXiv:1409.8211v2 fatcat:rgrxmkwbqjhvxh2rdqm6d7iyxe
« Previous Showing results 1 — 15 out of 3,551 results