System for recognizing or counting spoken itemized expressions

A method to recognize a spoken itemized expression comprising the following steps:

• a) identifying at least one segment indicative of speech followed by an adjacent segment indicative of a pause in said voiced speech and measuring a duration of the speech segment;
• and
• b) recognizing the speech segment as a spoken itemized expression when the measured duration of the speech segment is within a selected range of values;
• and
• c) recognizing an accented portion of the speech segment.

A method to recognize a spoken itemized expression in a speech segment, comprising the following steps:

• a) identifying at least one higher amplitude segment in the speech segment;
• b) identifying at least one subsequent adjacent lower amplitude segment in the speech segment;
• c) measuring the duration of the higher amplitude segment and the lower amplitude segment;
• d) determining a ratio initiative of comparative duration of the higher and lower amplitude segments;
• e) recognizing the higher amplitude segment as a spoken itemized expression when the ratio is within a selected range of values;
• and
• f) recognizing an accented portion of the higher amplitude segment.

A speech categorization system for counting portions of a voice message comprising at least one higher amplitude segment and at least one lower amplitude segment, at least some of said higher amplitude segments indicative of voiced speech, at least some of said lower amplitude segments indicative of periods of non-speech between adjacent voiced speech, said system comprising:

• a timer operative to generate first and second measured durations, wherein the first measured durations vary as a function of duration of selected higher amplitude segments, and wherein the second measured durations vary as a function of selected lower amplitude segments;
• means for recognizing and counting an accented portion of at least one of the higher amplitude segments;
• means for classifying a higher amplitude segment in a first category when the first and second measured durations corresponding to the higher amplitude segment and an adjacent and subsequent lower amplitude segment satisfy a classification test;
• and
• a counter operative to count the number of higher amplitude segments classified in the first category.

A speech categorization system for categorizing portions of a voice message comprising at least one higher amplitude segment and at least one lower amplitude segment,.at least some of said higher amplitude segments indicative of voiced speech, at least some of said lower amplitude segments indicative of periods of non-speech between adjacent voiced speech, said system comprising:

• a timer operative to generate first and second measured durations, wherein the first measured durations vary as a function of duration of selected higher amplitude segments, and wherein the second measured durations vary as a function of selected lower amplitude segments;
• means for recognizing an accented portion of at least one of the higher amplitude segments;
• and
• means for classifying a higher amplitude segment as a spoken itemized expression based on when the first and second measured durations for the higher amplitude segment and the adjacent subsequent lower amplitude segment satisfy a classification test.

A speech categorization system for categorizing portions of a voice message comprising at least one higher amplitude segment and at least one lower amplitude segment, at least some of said higher amplitude segments indicative of voiced speech, at least some of said lower amplitude segments indicative of periods of non-speech between adjacent voiced speech, said system comprising:

• means for developing a classification parameter indicative of comparative duration of at least some of the higher amplitude segments with respect to adjacent lower amplitude segments;
• means for classifying the higher amplitude segments into at least two categories, wherein one of said categories is indicative of a spoken itemized expression, said classifying means applying at least one classification test, said test operative to place one of the higher amplitude segments in a first category when the respective classification parameter is greater than a threshold value and to place said one of the higher amplitude segments in a second category when the respective classification parameter is less than the threshold value;
• means for recognizing and counting an accented portion of at least one of the higher amplitude segments;
• and
• a counter that compares the higher amplitude segments in one of the two categories to characterize the voice message.