From WER and RIL to MER and WIL: improved evaluation measures for connected speech recognition

By A. Morris, V. Maier, Phil D. Green. Interspeech, 2004

358 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Using IPA-Based Tacotron for Data Efficient Cross-Lingual Speaker Adaptation and Pronunciation Enhancement (2020, methodology): “Then, we computed the WER and MER [15] for the ASR model’s output, the original file, and the generated speech from a model in one of the transfer learning models.”
Human Evaluation and Correlation with Automatic Metrics in Consultation Note Generation (2022, neutral): “They include: Levenshtein edit distance (Levenshtein et al., 1966), WER (Su et al., 1992), MER and WIL (Morris et al., 2004).”
EvalCrafter: Benchmarking and Evaluating Large Video Generation Models (2023, methodology): “Then, we calculate Word Error Rate (WER) [30], Normalized Edit Distance (NED) [51], Character Error Rate (CER) [38], and finally we average these three score to get the OCR-Score.”