WSJCAM-0

The WSJCAM-0 dataset is an adult speech dataset used to train keyword spotting (KWS) systems, particularly in the context of evaluating their performance on children's speech.

Papers using WSJCAM-0 (3)

Zero-Shot KWS for Children's Speech using Layer-Wise Features from SSL Models2025 · 3 citesCan Layer-wise SSL Features Improve Zero-Shot ASR Performance for Children's Speech?2025 · 2 citesUnsupervised data selection for Speech Recognition with contrastive loss ratios2022