PENERAPAN VOICE RECOGNITION DALAM VERIVIKASI AKUN DENGAN METODE STATISTIK

22094690, Ratna Diana (2015) PENERAPAN VOICE RECOGNITION DALAM VERIVIKASI AKUN DENGAN METODE STATISTIK. Final Year Projects (S1) thesis, Universitas Kristen Duta Wacana.

	Text (Skripsi Informatika) 22094690_bab1_bab5_daftarpustaka.pdf Download (2MB)
	Text (Skripsi Informatika) 22094690_bab2-sd-bab4_lampiran.pdf Restricted to Registered users only Download (4MB) | Request a copy

Abstract

Securing data on electronic devices is an important action which must be undertaken when protecting the secrecy of said data. One common way of letters, numbers, or a combination of the two. Now, people can also use voice protection for their data; this provides stronger protection, because there is less of possibility that another person will be able to recreate it using the same voice as password creator. There are many ways to implement voice-based data production. One of the ways to implement voice-based data production is using the statistics method, through which the program will perform calculations on the voice of a person to determine whether or not that person has the right to access data. This is done by extracting three characteristics: the mean, standard deviations, and average deviations of a person’s voice. Research indicates that, by using the statistic method, the success ratio of identification decreases as the number of users increases. Furthermore, the number of voice clips (voice clipping is completed before extraction) will influence the rate of successful identification, to the point that it is necessary to conduct multiple tests to find the appropriate number of voice clips. The correct number of voice clips will optimize the identification rate; an increasing number of voice clips will necessitate an increasingly large threshold.

Actions (login required)

View Item