초록 ▼

A method of determining a dog's emotions from its voice with objective supports. The invention follows the procedures of converting dog's voices into electrical audio signals, extracting characteristics in a time to frequency component relation map of the audio signals as a input voice pattern, stor

A method of determining a dog's emotions from its voice with objective supports. The invention follows the procedures of converting dog's voices into electrical audio signals, extracting characteristics in a time to frequency component relation map of the audio signals as a input voice pattern, storing in advance in memory reference voice patterns for various emotions that respectively represent characteristics of time to frequency component relation maps, comparing the input voice pattern with the reference voice patterns, and determining what a dog feels by declaring emotion of the particular reference voice pattern showing the highest correlation with the input voice pattern as a result of the comparison, and emotions represented by the reference voice patterns include “loneliness”, “frustration”, “aggressiveness”, “assertiveness”, “happiness”, and “wistfulness”, and hence, the present invention attains an effective way of determining dog's emotions with objective supports.

대표청구항 ▼

1. An apparatus of determining dog's emotions by vocal analysis of dog's voice patterns, comprising means for converting dog's voices into electrical audio signals,means for extracting characteristics in a time to frequency component relation map of the audio signals as an input voice pattern,means

1. An apparatus of determining dog's emotions by vocal analysis of dog's voice patterns, comprising means for converting dog's voices into electrical audio signals,means for extracting characteristics in a time to frequency component relation map of the audio signals as an input voice pattern,means for storing reference voice patterns that represent various dog's emotions expressed by its voice in a form of the time to frequency component relation map, means for comparing the input voice pattern with the reference voice patterns, andmeans for determining how the dog feels by choosing the pattern most closely matched to the input voice pattern;wherein the apparatus being characterized in that the reference voice patterns includea reference voice pattern expressing loneliness that is made up of significant frequency components around 5000 Hz without frequency components of 3000 Hz or below but lacks harmonics and lasts for 0.2 to 0.3 second,a reference voice pattern expressing frustration that is made up of a sub-pattern for 0.3 to 1 second of fundamental tones ranging from 160 to 240 Hz with harmonics up to 1500 Hz and a subsequent sub-pattern of 250 to 8000 Hz without definite fundamental tones and harmonics but is conspicuous for frequency components around 1000 Hz,a reference voice pattern expressing aggressiveness that is made up of a sub-pattern ranging from 250 to 8000 Hz without definite fundamental tones and harmonics and conspicuous for frequency components around 1000 Hz and a subsequent sub-pattern for 0.8 to 1.5 second of fundamental tones ranging from 240 to 360 Hz with definite harmonics up to 1500 Hz and reaching 8000 Hz,a reference voice pattern expressing assertiveness that is made up of frequency components ranging from 250 to 8000 Hz without definite fundamental tones and harmonics and is conspicuous for frequency components around 1000 Hz, 2000 Hz and 5000 Hz, respectively,a reference voice pattern expressing happiness including a sub-pattern of frequency components ranging from 250 to 8000 Hz without definite fundamental tones and harmonics and with conspicuous frequency components around 1000 Hz and a subsequent sub-pattern of fundamental tones ranging from 200 to 300 Hz with harmonics up to 1500 Hz, anda reference voice pattern expressing wistfulness that is made up of fundamental tones ranging from 250 to 500 Hz with harmonics up to 8000 Hz and with frequency of the fundamental tones being varied. 2. A method for determining dog's emotions by vocal analysis of dog's voice patterns, comprising the steps ofconverting dog's voices into electrical audio signals, extracting characteristics in a time to frequency component relation map of the audio signals as an input voice pattern,storing reference voice patterns that represent various dog's emotions expressed by its voice in a form of the time to frequency component relation map,comparing the input voice pattern with the reference voice patterns, and determining how the dog feels by choosing the pattern most closely matched to the input voice pattern;the method being characterized in that the reference voice patterns includea reference voice pattern expressing loneliness that is made up of significant frequency components around 5000 Hz without frequency components of 3000 Hz or below but lacks harmonics and lasts for 0.2 to 0.3 second,a reference voice pattern expressing frustration that is made up of a sub-pattern for 0.3 to 1 second of fundamental tones ranging from 160 to 240 Hz with harmonics up to 1500 Hz and a subsequent sub-pattern of 250 to 8000 Hz without definite fundamental tones and harmonics but is conspicuous for frequency components around 1000 Hz,a reference voice pattern expressing aggressiveness that is made up of a sub-pattern ranging from 250 to 8000 Hz without definite fundamental tones and harmonics and is conspicuous for frequency components around 1000 Hz and a subsequent sub-pattern for 0.8 to 1.5 second of fundamental tones rangin g from 240 to 360 Hz with harmonics definite up to 1500 Hz and reaching 8000 Hz,a reference voice pattern expressing assertiveness that is made up of frequency components ranging from 250 to 8000 Hz without definite fundamental tones and harmonics and is conspicuous for frequency components around 1000 Hz, 2000 Hz and 5000 Hz, respectively,a reference voice pattern expressing happiness including a sub-pattern of frequency components ranging from 250 to 8000 Hz without definite fundamental tones and harmonics and with conspicuous frequency components around 1000 Hz and a subsequent sub-pattern of fundamental tones ranging from 200 to 300 Hz with harmonics up to 1500 Hz, anda reference voice pattern expressing wistfulness that is made up of fundamental tones ranging from 250 to 500 Hz with harmonics up to 8000 Hz and with frequency of the fundamental tones being varied.