VoxVietnam: a Large-Scale Multi-Genre Dataset for Vietnamese Speaker Recognition

2024-12-31

AI Summary

The absence of large-scale, multi-genre (e.g., news, interviews, podcasts, telephone) benchmark datasets hinders robust speaker recognition for Vietnamese, a low-resource language. Method: We introduce VoxVietnam—the first open-source, multi-genre Vietnamese speaker recognition dataset—comprising 187,000 utterances from 1,406 speakers. We propose a fully automated, reproducible end-to-end pipeline integrating web crawling, ASR-based transcription, speaker diarization via spectral clustering, and rigorous quality filtering. Contribution/Results: We systematically model and empirically validate the substantial impact of genre variation on speaker verification performance: models trained on a single genre suffer a 12.6% increase in equal error rate (EER) when evaluated cross-genre. Leveraging VoxVietnam for training reduces cross-genre EER by 38.2%, significantly improving generalization. This work establishes a critical data foundation and methodological paradigm for robust speaker recognition in low-resource languages.

Technology Category

• Natural Language Processing: Speech
• Data Mining & Knowledge Management: Mining of Visual, Multimedia & Multimodal Data
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

Recent research in speaker recognition aims to address vulnerabilities due to variations between enrolment and test utterances, particularly in the multi-genre phenomenon where the utterances are in different speech genres. Previous resources for Vietnamese speaker recognition are either limited in size or do not focus on genre diversity, leaving studies in multi-genre effects unexplored. This paper introduces VoxVietnam, the first multi-genre dataset for Vietnamese speaker recognition with over 187,000 utterances from 1,406 speakers and an automated pipeline to construct a dataset on a large scale from public sources. Our experiments show the challenges posed by the multi-genre phenomenon to models trained on a single-genre dataset, and demonstrate a significant increase in performance upon incorporating the VoxVietnam into the training process. Our experiments are conducted to study the challenges of the multi-genre phenomenon in speaker recognition and the performance gain when the proposed dataset is used for multi-genre training.

Problem

• Vietnamese speaker recognition
• Diverse speech types
• Performance evaluation

Innovation

• VoxVietnam
• Multigenre Vietnamese Speech Recognition
• Cross-genre Recognition Performance