Convolutional recurrent neural networks for music classification

By Keunwoo Choi, György Fazekas, Mark B. Sandler, Kyunghyun Cho. IEEE International Conference on Acoustics, Speech, and Signal Processing, 2016

536 citations recorded by the source corpus.

We introduce a convolutional recurrent neural network (CRNN) for music tagging. CRNNs take advantage of convolutional neural networks (CNNs) for local feature extraction and recurrent neural networks for temporal summarisation of the extracted features. We compare CRNN with three CNN structures that have been used for music tagging while controlling the number of parameters with respect to their performance and training time per sample. Overall, we found that CRNNs show a strong performance with respect to the number of parameter and training time, indicating the effectiveness of its hybrid structure in music feature extraction and feature summarisation.

Subfields: Computer Vision and Pattern Recognition, Signal Processing

View Paper: https://www.semanticscholar.org/paper/5f20da406e874ee0aca065ab4bf6024e8f1d0d4c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Signal Processing

Citations in the corpus

Sample-level Deep Convolutional Neural Networks for Music Auto-tagging Using Raw Waveforms (2017, background): “The best results were obtained when m was 3 and n was 9.”
MidiNet: A Convolutional Generative Adversarial Network for Symbolic-Domain Music Generation (2017, background): “We are also interested in using ideas of reinforcement learning [29] to incorporate principles of music theory [18], and to take input from music information retrieval models such as genre recognition [6] and emotion recognition [34].”
On Translation Invariance in CNNs: Convolutional Layers Can Exploit Absolute Spatial Location (2020, methodology): “The CNN is now even used in other modalities such as speech [1, 58, 73], audio [15, 40, 82], text [14, 20, 56], graphs [11, 21, 84], etc .”
Convolutional Recurrent Neural Networks for Polyphonic Sound Event Detection (2017, unknown): “Similar approaches combining CNNs and RNNs have been presented recently in ASR [21], [28], [29] and music classiﬁcation [30].”
Vggsound: A Large-Scale Audio-Visual Dataset (2020, background): “More recently, the success of deep learning has motivated approaches based on CNNs [5, 27] or RNNs [28, 29, 30].”
Evaluation of CNN-based Automatic Music Tagging Models (2020, methodology): “Convolutional recurrent neural network (CRNN) [6] uses Mel spectrogram inputs.”
Multi-Level and Multi-Scale Feature Aggregation Using Pretrained Convolutional Neural Networks for Music Auto-Tagging (2017, unknown): “…architecture, which is designed considering different levels of abstractions in Bag of multi-scaled features [8] 0.898 - - 1D CNN [2] 0.8815 - - Transfer learning [13] 0.88 - - Persistent CNN [22] 0.9013 - - Time-Frequency CNN [23] 0.9007 - - 2D CNN [4] 0.894 0.851 - CRNN [24] - 0.862 - 2D CNN [”