Multi-label Ferns for Efficient Recognition of Musical Instruments in Recordings

TL;DR

This paper introduces multi-label ferns for efficient and accurate classification of musical instruments in audio recordings, demonstrating faster processing, higher accuracy, and smaller models compared to binary random ferns.

Contribution

The paper presents a novel multi-label fern method that improves classification speed, accuracy, and model compactness over existing binary fern approaches.

Findings

Faster classification than binary random ferns.

Higher F-score in instrument recognition.

Significant reduction in model size.

Abstract

In this paper we introduce multi-label ferns, and apply this technique for automatic classification of musical instruments in audio recordings. We compare the performance of our proposed method to a set of binary random ferns, using jazz recordings as input data. Our main result is obtaining much faster classification and higher F-score. We also achieve substantial reduction of the model size.