Speech-Driven End-to-End Language Discrimination towards Chinese Dialects

In three linesPaper presents speech-driven approach for Chinese dialect discrimination. Combines MFCC features, HMM-DNN speech recognition model, attention mechanism and CNN. Evaluation on two benchmark Chinese dialect corpora shows improvement over state-of-the-art methods.

Summary generated by Claude — human-verified