New analysis improves children's speech recognition with F-TDNN

Researchers have analyzed fine-tuning strategies for automatic speech recognition (ASR) systems, specifically focusing on low-resource scenarios involving children's speech. The study investigated various acoustic features and their impact on different acoustic models, finding that pitch features significantly improved recognition performance for dysarthric speech. By systematically examining the TORGO database with a Factorized Time Delay Neural Network (F-TDNN) model, the team achieved relative improvements of 4.65% in isolated word recognition and 4.63% in sentence recognition. AI

IMPACT This research could lead to more accurate speech recognition systems for children, particularly those with speech impairments.

RANK_REASON The cluster contains an academic paper detailing a new analysis and methodology for speech recognition. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Factorized Time Delay Neural Network
• F-TDNN
• TORGO database

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →