UX and Machine Learning – Preprocessing of Audiovisual Data Using Computer Vision to Recognize UI Elements

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60460709%3A41110%2F23%3A96163" target="_blank" >RIV/60460709:41110/23:96163 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.7160/aol.2023.150304" target="_blank" >https://doi.org/10.7160/aol.2023.150304</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.7160/aol.2023.150304" target="_blank" >10.7160/aol.2023.150304</a>

Jazyk výsledku

angličtina

Název v původním jazyce

UX and Machine Learning – Preprocessing of Audiovisual Data Using Computer Vision to Recognize UI Elements

Popis výsledku v původním jazyce

This study explores the convergence of user experience (UX) and machine learning, particularly employing computer vision techniques to preprocess audiovisual data to detect user interface (UI) elements. With an emphasis on usability testing, the study introduces a novel approach for recognizing changes in UI screens within video recordings. The methodology involves a sequence of steps, including form prototype creation, laboratory experiments, data analysis, and computer vision tasks. The future aim is to automate the evaluation of user behavior during UX testing. This innovative approach is relevant to the agricultural domain, where specialized applications for precision agriculture, subsidy requests, and production reporting demand streamlined usability. The research introduces a frame extraction algorithm that identifies screen changes by analyzing pixel differences between consecutive frames. Additionally, the study employs YOLOv7, an efficient object detection model, to identify UI elements within the video frames. Results showcase successful screen change detection with minimal false negatives and acceptable false positives, showcasing the potential for enhanced automation in UX testing. The study’s implications lie in simplifying analysis processes, enhancing insights for design decisions, and fostering user-centric advancements in diverse sectors, including precision agriculture.

Název v anglickém jazyce

UX and Machine Learning – Preprocessing of Audiovisual Data Using Computer Vision to Recognize UI Elements

Popis výsledku anglicky

This study explores the convergence of user experience (UX) and machine learning, particularly employing computer vision techniques to preprocess audiovisual data to detect user interface (UI) elements. With an emphasis on usability testing, the study introduces a novel approach for recognizing changes in UI screens within video recordings. The methodology involves a sequence of steps, including form prototype creation, laboratory experiments, data analysis, and computer vision tasks. The future aim is to automate the evaluation of user behavior during UX testing. This innovative approach is relevant to the agricultural domain, where specialized applications for precision agriculture, subsidy requests, and production reporting demand streamlined usability. The research introduces a frame extraction algorithm that identifies screen changes by analyzing pixel differences between consecutive frames. Additionally, the study employs YOLOv7, an efficient object detection model, to identify UI elements within the video frames. Results showcase successful screen change detection with minimal false negatives and acceptable false positives, showcasing the potential for enhanced automation in UX testing. The study’s implications lie in simplifying analysis processes, enhancing insights for design decisions, and fostering user-centric advancements in diverse sectors, including precision agriculture.

Druh

J_SC - Článek v periodiku v databázi SCOPUS

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

AGRIS on-line Papers in Economics and Informatics

ISSN

1804-1930

e-ISSN

1804-1930

Svazek periodika

15

Číslo periodika v rámci svazku

03/2023

Stát vydavatele periodika

CZ - Česká republika

Počet stran výsledku

10

Strana od-do

35-44

Kód UT WoS článku

—

EID výsledku v databázi Scopus

2-s2.0-85173683350