VoxCeleb: a large-scale speaker identification dataset
This provides a valuable resource for researchers in speaker identification, addressing a bottleneck in the field, though it is incremental as it builds on existing methods for data collection.
The authors tackled the lack of large-scale, real-world datasets for speaker identification by creating VoxCeleb, a dataset with hundreds of thousands of utterances from over 1,000 celebrities, and established baseline performance showing CNN-based architectures achieve the best results.
Most existing datasets for speaker identification contain samples obtained under quite constrained conditions, and are usually hand-annotated, hence limited in size. The goal of this paper is to generate a large scale text-independent speaker identification dataset collected 'in the wild'. We make two contributions. First, we propose a fully automated pipeline based on computer vision techniques to create the dataset from open-source media. Our pipeline involves obtaining videos from YouTube; performing active speaker verification using a two-stream synchronization Convolutional Neural Network (CNN), and confirming the identity of the speaker using CNN based facial recognition. We use this pipeline to curate VoxCeleb which contains hundreds of thousands of 'real world' utterances for over 1,000 celebrities. Our second contribution is to apply and compare various state of the art speaker identification techniques on our dataset to establish baseline performance. We show that a CNN based architecture obtains the best performance for both identification and verification.