Samodejno razpoznavanje govora v bolnišnicah

FRANTAR, JURE (avtor), Rozman, Robert (mentor)

Zaradi lažje organizacije dela ter večje preglednosti nad delom zaposlenih v medicinskih ustanovah že nekaj časa vgrajujemo Android sobne terminale, občutljive na dotik. Slednji imajo nameščene prilagojene aplikacije, ki služijo kot pomoč pri delu v medicinskih ustanovah. Njihova glavna naloga je prikaz aktivnih dogodkov, informacij ter beleženje storitev, ki jih zaposleni opravijo bolnikom. Zaposleni pa med opravljanjem storitev velikokrat potrebujejo pomoč dodatne sestre ali zdravnika. Zaradi lažjega opravljanja storitev brez prekinitev dela ob klicu dodatnega zaposlenega smo funkcionalnost terminalov nadgradili z razpoznavanjem govora, ki ga podrobneje opisujemo v tem delu. Za zaznavanje ključne besede smo uporabili orodje CMUSphinx, za razpoznavanje ukazov uporabnikov pa orodje Android.SpeechRecognizer. Postopek za samodejno tvorbo govornih odzivov na ukaze uporabnikov pa smo zasnovali na orodju TextToSpeech. Poleg zaposlenih pa lahko terminal uporabljajo tudi bolniki. V angleškem jeziku lahko izrazijo svoje želje zaposlenim, kateri lahko pridejo v sobo že ustrezno pripravljeni. Z dodano funkcionalnostjo smo poenostavili delo zaposlenih in izboljšali delovne procese v zdravstvenih ustanovah.

• Jezik: Slovenski jezik
• Ključne besede: sobni terminal na dotik, mobilna aplikacija NurseApp, razpoznavanje govora, medicina
• Vrsta gradiva: Diplomsko delo/naloga
• Organizacija: FRI - Fakulteta za računalništvo in informatiko
• Leto izida: 2018
• PID: 20.500.12556/RUL-105267
• Datum objave v RUL: 16.11.2018

FRANTAR, JURE, 2018, Samodejno razpoznavanje govora v bolnišnicah [na spletu]. Diplomsko delo. [Dostopano 31 marec 2025]. Pridobljeno s: https://repozitorij.uni-lj.si/IzpisGradiva.php?lang=slv&id=105267

Sekundarni jezik

• Jezik: Angleški jezik
• Naslov: Automatic Speech Recognition in Hospitals

Izvleček

To achieve the easier work organization and greater transparency over the work of medical staff, we have been installing touch-sensitive Android terminals in medical institutions for quite some time. The terminals have installed customized applications that serve as an assistance to the working process. Their main task is to present actual events, information and record the services that employees provide to patients. However, when providing services, employees often need the help of an additional employee. In order to facilitate the provision of services without the interruption due to the call of an additional employee, the functionality of the terminals was upgraded with the speech recognition, which is described in more detail in this work. We have based our work on the CMUSphinx tool to identify the spoken keywords and the Android.SpeechRecognizer tool to recognize commands and random speech of terminal users. We have implemented the automatic spoken responses to user commands with the TextToSpeech tool. In addition to employees, the terminals can also be used by patients. They can express their wishes in the spoken English language, and employees can come to the room already well prepared. With added functionality, we have simplified the work of employees and therefore improved work processes in medical institutions.

• Ključne besede: room touch terminals, mobile application NurseApp, speech recognition, medicine