Segment Aggregation for short utterances speaker verification using raw waveforms
This addresses the problem of reduced accuracy in speaker verification systems for short-duration utterances, which is an incremental improvement for applications like voice authentication.
The paper tackles performance degradation in speaker verification for short utterances by proposing a segment aggregation method that segments input utterances, aggregates embeddings, and trains them simultaneously, achieving a 45.37% relative improvement over the baseline with 1-second test utterances.
Most studies on speaker verification systems focus on long-duration utterances, which are composed of sufficient phonetic information. However, the performances of these systems are known to degrade when short-duration utterances are inputted due to the lack of phonetic information as compared to the long utterances. In this paper, we propose a method that compensates for the performance degradation of speaker verification for short utterances, referred to as "segment aggregation". The proposed method adopts an ensemble-based design to improve the stability and accuracy of speaker verification systems. The proposed method segments an input utterance into several short utterances and then aggregates the segment embeddings extracted from the segmented inputs to compose a speaker embedding. Then, this method simultaneously trains the segment embeddings and the aggregated speaker embedding. In addition, we also modified the teacher-student learning method for the proposed method. Experimental results on different input duration using the VoxCeleb1 test set demonstrate that the proposed technique improves speaker verification performance by about 45.37% relatively compared to the baseline system with 1-second test utterance condition.