Audio Scene Classification with Deep Recurrent Neural Networks

TL;DR

This paper presents a deep recurrent neural network approach for audio scene classification that transforms audio into label embeddings and achieves state-of-the-art accuracy on a large dataset.

Contribution

The work introduces a novel sequence-to-label classification method using deep GRUs and high-level label tree embeddings for audio scene recognition.

Findings

Achieved 97.7% F1-score on LITIS Rouen dataset.

Reduced classification error by 35.3% compared to previous best.

Demonstrated efficiency and effectiveness of RNNs for audio scene classification.

Abstract

We introduce in this work an efficient approach for audio scene classification using deep recurrent neural networks. An audio scene is firstly transformed into a sequence of high-level label tree embedding feature vectors. The vector sequence is then divided into multiple subsequences on which a deep GRU-based recurrent neural network is trained for sequence-to-label classification. The global predicted label for the entire sequence is finally obtained via aggregation of subsequence classification outputs. We will show that our approach obtains an F1-score of 97.7% on the LITIS Rouen dataset, which is the largest dataset publicly available for the task. Compared to the best previously reported result on the dataset, our approach is able to reduce the relative classification error by 35.3%.

Figures

Full text

See pages 1-last of main.pdf