Hao-Teng Fan

Learn More
The conventional NMF-based speech enhancement algorithm analyzes the magnitude spectrograms of both clean speech and noise in the training data via NMF and estimates a set of spectral basis vectors. These basis vectors are used to span a space to approximate the magnitude spectrogram of the noise-corrupted testing utterances. Finally, the components(More)
摘要 在本論文裡,我們提出了一種藉由線性估測編碼來強化語音辨識中特徵之抗噪 性的新方法,在此方法中,根據線性估測編碼技術,將語音倒頻譜特徵時間序 列分解出估測誤差成分後,將此估測誤差成分從原特徵序列扣除,所得的新特 徵序列,相對於原始特徵序列而言,發現具有更佳的雜訊強健性,在 Aurora-2 此包含各類雜訊之數字語料庫的實驗環境下,經過各種預強健化處理之倒頻譜 語音特徵,再進一步藉由我們所提之新方法處理後,都能得到更佳的辨識效能, 且在線性估測階數很低的情況下,就可有效提升辨識率,顯示了我們可以高效 率地執行實現所提之新技術。 關鍵詞:線性估測編碼、特徵時間序列、雜訊強健性。 Abstract In this paper, we present a novel method to extract(More)
—In this paper, we present a novel approach to enhancing the speech features in the modulation spectrum for better recognition performance in noise-corrupted environments. In the presented approach, termed modulation spectrum power-law expansion (MSPLE), the speech feature temporal stream is first pre-processed by some statistics compensation technique,(More)