Technical Paper
A New U-Net Speech Enhancement Framework Based on Correlation Characteristics of Speech
2024-04-09
2024-01-2015
As a key component of in-vehicle intelligent voice technology, speech enhancement can extract clean speech signals contaminated by environmental noise to improve the quality and intelligibility of speech perception. It has broad applications in human-vehicle interaction, in-car communication, and in-car cinema. Although some end-to-end time-domain-based speech enhancement methods have been proposed, few models are designed based on the characteristics of speech signals. In this paper, we propose a new U-Net based speech enhancement method that utilizes the temporal correlation of speech signals to reconstruct higher-quality and more intelligible clean speech.