Designing the Intelligent System Detecting a Sense of Wonder in English Speech Signal Using Fuzzy-Nervous Inference-Adaptive system (ANFIS)

Sakine Tashakori, Salman Haghighat


The purpose of this research is to design an intelligent diagnostic system for detecting a sense of wonder in English speech signal using Fuzzy-Nervous Inference-Adaptive system (ANFIS). For English, the recognition of some surprise feelings such as anger, grief, joy and hatred has been made, but due to the difficulty of creating a speech database in a state of wonder and a shortage of resources in this case, even in other languages, so far, no sense of wonder has been detected in the English speech. In the absence of a suitable database in English for the identification of emotions, at first, a wonder-neutral database (without feeling) was created in Persian, containing 30 sentences with a sense of surprise and neutrality. Then, LPC coefficients and frequency characteristics of speech signals such as maximum, minimum, middle and mean (obtained by FFT) were extracted. Finally, the neuro-fuzzy adaptive network (ANFIS) was used to create a sense of wonder with an average accuracy of about 94.93%.


Emotion Detection; English Speech Signal; Fuzzy-Nervous Inference-Adaptive System (ANFIS)

Full Text:


(C) 2010-2022 EduSoft