Download now Free registration required
In this paper, the authors propose a new histogram equalization technique for feature compensation in speech recognition under noisy environments. The proposed approach combines a signal-to-noise-ratio - dependent feature reconstruction method and the class histogram equalization technique to effectively reduce the acoustic mismatch present in noisy speech features. Experimental results from the Aurora 2 task confirm the superiority of the proposed approach for acoustic feature compensation. The performance of speech recognizers degrades severely when they are employed in acoustically mismatched environments compared to training environments.
- Format: PDF
- Size: 254.77 KB