There are some commercial softwares to do this kind of job. As Robert has referred,Dragon Natural Speaking is one of those softwares. But this software is just to translate the voice into the text, it DO NOT differentiate the persons who say the same word.(Is it true?).
"Voice recognition or speaker recognition is a related process that attempts to identify the person speaking, as opposed to what is being said."
For biological password, if everybody has unique voice which can be differentiated by the software, not only the software differentiate the words the user said,but also it can capture the essence of the voice which is unquely belonged the specific person.
I will do the survey to try to find whether we can use the voice to identify a person like using the fingerprint .
What the project will do is not to translate the voice to text as the most commercial softwares do, in my opinon, we just need analyze the voice and word (password which is not very long) set by the user, encrypted and stored it in the database, the next time the user wants to log into the machine, he needs to speak that words again to the microphone. That is it.
But it may be quite difficult,because there are two things,actually it includes the content and tone. The easy part for this, the software should "learn the word and tone ,remember it, then it can recognize it." For the current commercial software, it does not have the opportunity to learn all the words, it is dependent on its database. It should compare the words with its database to find the match, the database should be large enough.
For voice password, we just need to store the password in the database, which can be set up by the user. It simplifyies the problem. The difficult part is how to grasp the password's essence, how can grasp the unique feature of the voice of the specific person.
No comments:
Post a Comment