The automatic recognition of pathological speech, particularly from children with any articulatory impairment, is a challenging task due to various reasons. The lack of available domain specific data is one such obstacle that hinders its usage for different speech-based applications targeting pathological speakers. In line with the challenge, in this work, we investigate a few data augmentation techniques to simulate training data for improving the children speech recognition considering the case of cleft lip and palate (CLP) speech. The augmentation techniques explored in this study, include vocal tract length perturbation (VTLP), reverberation, speaking rate, pitch modification, and speech feature modification using cycle consistent adversarial networks (CycleGAN). Our study finds that the data augmentation methods significantly improve the CLP speech recognition performance, which is more evident when we used feature modification using CycleGAN, VTLP and reverberation based metho...

— It is common for children who are born with a cleft lip and palate issues. Therefore they have speech problems at some time in their lives. Over half of them will require speech therapy at some point during childhood. The speech language pathologist has to access child's speech production and language development and make appropriate therapy recommendation. Once the palate has been repaired, child may be able to learn more consonant sounds and say more words but speech may still be delayed and difficulties in making certain sound. Therefore those children are treated by speech therapist clinic. The children with parents has to come to clinic and get the speech therapy. It is a continues process. During the therapy period, children's' speech may be understandable, but some sound may be distorted. To address the issue a biometric solution can be implemented. Speech recognition is used to identify the speak by the children. The speech recognition aims at understanding and comprehending WHAT is spoken. In this research, automated speech recognition system is proposed to identifying certain sound from the child then extract this sound and match with correct sound. Final stage, the comparison and decision making are taken place.