Priyaranjan Pattnayak
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
NLP×1
Research Timeline
2026
SN-WER: Script-Normalized WER for Multi-Script Indic ASR Evaluation
The paper introduces Script-Normalized WER (SN-WER), a novel evaluation metric that transliterates ASR transcripts into a canonical script to accurately measure speech recognition performance across different scripts, showing significant reduction in artificially inflated error rates.
Highlighted terms show continued research focus across papers