Affordable Access

Speaker Identification of Whispered Speech with Perceptible Mood

Authors
Publisher
ACADEMY PUBLISHER
Publication Date
Keywords
  • Speaker Identification
  • Whispered Speech
  • Perceptive Mood
  • Latent Factor Analysis
Disciplines
  • Computer Science

Abstract

Special attention has been paid to SI (speaker identification) of whispered speech with perceptible mood. The database is developed for SI of whispers with different purposes. To prove the efficiency of delivering the speaker’s feelings, the perceptual test is carried out on the whispered speech, and the A-V (Arousal-Valance) space is introduced to judge the speaker’s status. The LFA (Latent Factor Analysis) algorithm, following the classification of the speaker’s mood, is introduced in detail and proved to be efficient. The proposed SI scheme with modified spectrum parameters provides much more satisfactory accuracy than the baseline system.

There are no comments yet on this publication. Be the first to share your thoughts.