Content-based Representations of audio using Siamese neural networks

摘要

In this paper, we focus on the problem of content-based retrieval for audio,which aims to retrieve all semantically similar audio recordings for a givenaudio clip query. We propose a novel approach which encodes the audio into avector representation using Siamese Neural Networks. The goal is to obtain anencoding similar for files belonging to the same audio class, thus allowingretrieval of semantically similar audio. We used two similarity measures,Cosine similarity and Euclidean distance, to show that our method is effectivein retrieving files similar in audio content. Our results indicate that ourneural network-based approach is able to retrieve files similar in content andsemantics.