Enhancing Children’s Short Utterance Based ASV Using Data Augmentation Techniques and Feature Concatenation Approach

Abstract

The task of developing an automatic speaker verification (ASV) system for children’s speech is a challenging one due to a number of reasons. The dearth of domain-specific data is one among them. The challenge further intensifies with the introduction of short utterances of speech, a relatively unexplored domain in the case of children’s ASV. To circumvent the issue arising due to data scarcity, the work in this paper extensively explores various in-domain and out-of-domain data augmentation techniques. A data augmentation approach is proposed that encompasses both in-domain and out-of-domain data augmentation techniques. The out-of-domain data used are from adult speakers which are known to have acoustic attributes in stark contrast to child speakers. Consequently, various techniques like prosody modification, formant modification and voice-conversion are employed in order to modify the adult acoustic features and render it acoustically similar to children’s speech prior to augmentation. The in-domain data augmentation approach, on the other hand, involved speed perturbation of children’s speech. The proposed data augmentation approach helps not only in increasing the amount of training data but also in effectively capturing the missing target attributes which helps in boosting the verification performance. A relative improvement of $43.91\%$ in equal error rate (EER) with respect to the baseline system is a testimony of it. Furthermore, the commonly used Mel-frequency cepstral coefficients (MFCC) average out the higher-frequency components due to the larger bandwidth of the filter-bank. Therefore, effective preservation of higher-frequency contents in children’s speech is another challenge which must be appropriately tackled for the development of a reliable and robust children’stion techniques and Feature Concatenation A ASV system. The feature concatenation of MFCC and IMFCC is carried out with the sole intention of effectively preserving the higher-frequency contents in the children’s speech data. The feature concatenation approach, when combined with proposed data augmentation, helps in further improvement of the verification performance and results in an overall relative reduction of $48.51\%$ for equal error rate.