LibriPhrase Hard

The 'LibriPhrase Hard' dataset/benchmark contains challenging phrases designed to evaluate the performance of keyword spotting systems, particularly in distinguishing confusable words and handling variations in speaker pronunciations.

Papers using LibriPhrase Hard (3)

Effective User-defined Keyword Spotting with Dual-stage Matching, Multi-modal Enrollment, and Continual Adaptation2026Flexible Keyword Spotting based on Homogeneous Audio-Text Embedding2023 · 1 citesOpen vocabulary keyword spotting through transfer learning from speech synthesis2024 · 1 cites