Sterowanie pracą komputera za pomocą komend słownych

Brak streszczenia.

In 1968 Stanley Kubrick directed a legendary science-fiction movie “2001: A Space Odyssey”. The
main “character” in this famous production was HAL 9000 – a supercomputer with much higher
intelligence than human. The machine was not only capable of carrying fluent conversations with
people, but also could easily read speech following mouth motion.
Although the contemporary automatic speech recognition systems are incomparable to HAL, this
branch of artificial intelligence is developing very fast, since it gives enormous possibilities in many
domains of science.
The aim of this research was to design, implement and deploy a speech recognition desktop
application providing the possibility to fully control the computer by means of human voice. It
should work in real time and have easy graphical user interface.
The assumption is that the system should recognize several isolated words. Each of these should be
associated with certain system action – normally serviced with a mouse, keyboard or another I/O
device. Since the most promising results in contemporary studies in speech recognition are
generated by Hidden Markov Models, it was decided to choose it as an implementation algorithm.
The modeling is based on Linear Predictive Coding applied to recorded words.
The application was named AudioBoss. The first part of this word originates from the assumption
that the application should be controlled by means of user's phonetic capabilities. The second part
represents the controlling functionality of the system.
The next Chapter of this dissertation will provide the Reader with information on the scientific
context of the issue. Later, a full description of the system design will be elaborated. This includes
software environment presentation, system architecture overview, signal acquisition and analysis
description, Hidden Markov Models creation and word recognition algorithm. The graphical user
interface and user manual will be described in Chapter 4. The operation of the system has been
evaluated experimentally and the obtained results have been presented in Chapter 5.