What question did this study set out to answer?

This research aims to create a new annotated dataset for environmental sound classification and evaluate classification accuracy.

October 13, 2015

Esc

View Full Paper

Ask AI

Bookmark

This research aims to create a new annotated dataset for environmental sound classification and evaluate classification accuracy.
Created a collection of 2000 annotated sound clips across 50 classes.
Compiled 250,000 unlabeled auditory excerpts from the Freesound project.
Evaluated human classification accuracy and compared it to baseline classifiers using specific audio features.
Human accuracy in sound classification was assessed and found effective based on the dataset.
Baseline classifiers showed varying performance metrics based on mel-frequency features and zero-crossing rate.

No takes yet. Share an insight, caveat, or question.

Randomized trial evaluates human and classifier accuracy in classifying environmental sounds, indicating dataset utility.

Karol J. Piczak (2015) studied this question.

View Full Paper

Ask AI

Bookmark