Hand sign language recognition with artificial intelligence : Using “You Only Look Once” (Yolo) model as a case
Bui, Hien (2022)
2022
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2022060113945
https://urn.fi/URN:NBN:fi:amk-2022060113945
Tiivistelmä
This thesis introduces information on sign language and the culture around it. Additionally, what was still lacking in the other existing research before and after Yolo is introduced including my unique contributions to this topic using the constructive research method.
In this thesis, the information about sign language and deaf-and-mute community was referenced from related books and research to provide the readers reliable insight about this new field. Furthermore, the technical knowledge was collected from Internet sources and scientific articles to provide an up-to-date and accurate information at the time this thesis was written. The data source that was used to train the Yolo model was manually collected from the YouTube videos to ensure its diversity.
For demonstration, a website application was created to evaluate the reliability of using the Yolo model for translating American sign language alphabets. Finally, the conclusion of this thesis work indicate that Yolo version 3 was not an optimal solution for sign language despite its outstanding speeds and efficiency in static object detection.
In this thesis, the information about sign language and deaf-and-mute community was referenced from related books and research to provide the readers reliable insight about this new field. Furthermore, the technical knowledge was collected from Internet sources and scientific articles to provide an up-to-date and accurate information at the time this thesis was written. The data source that was used to train the Yolo model was manually collected from the YouTube videos to ensure its diversity.
For demonstration, a website application was created to evaluate the reliability of using the Yolo model for translating American sign language alphabets. Finally, the conclusion of this thesis work indicate that Yolo version 3 was not an optimal solution for sign language despite its outstanding speeds and efficiency in static object detection.