Automatic speaker verification system for dysarthric speakers using prosodic features and out-of-domain data augmentation

•To the best of our knowledge, this is the first work dealing in detail the nuances of developing an ASV system for dysarthria patients.•Investigated prosody based measures and studied the effects of combining MFCC and prosodic features on an ASV system developed for dysrthria speakers. We have also investigated the role of duration modification based data augmentation to deal with the paucity of domain-specific speech data in order to develop effective ASV system.•Experimental comparison of x-vector and i-vector based speaker embeddings is also performed in the context of dysarthric speaker ASV.•Severity wise analysis of system performance was done to understand the effect of data augmentation along with MFCC and prosody features with variation in the disease severity of the dysarthria speakers.•The paper also compares the duration modification based data augmentation technique with other related methods and analyzed the performance of dysarthric speakers with different severity.