General-purpose Tagging of Freesound Audio with AudioSet Labels: Task Description, Dataset, and Baseline

TL;DR

This paper introduces a new audio tagging task using Freesound clips labeled with AudioSet categories, providing a dataset, task description, and baseline system to advance general-purpose audio recognition.

Contribution

It presents a comprehensive task setup, dataset, and baseline for audio tagging of Freesound content with AudioSet labels, facilitating future research.

Findings

Dataset and task description for Freesound audio tagging

Baseline system performance established

Encourages development of improved audio tagging models

Abstract

This paper describes Task 2 of the DCASE 2018 Challenge, titled "General-purpose audio tagging of Freesound content with AudioSet labels". This task was hosted on the Kaggle platform as "Freesound General-Purpose Audio Tagging Challenge". The goal of the task is to build an audio tagging system that can recognize the category of an audio clip from a subset of 41 diverse categories drawn from the AudioSet Ontology. We present the task, the dataset prepared for the competition, and a baseline system.