An integrated framework for user interface design optimization using real-time eye tracking analysis and machine learning

Natalia, Nila and Mohamed, Mazlan and Ab Ghani, Hadhrami (2025) An integrated framework for user interface design optimization using real-time eye tracking analysis and machine learning. Borneo Journal of Sciences and Technology (BJOST), 8 (1). pp. 12-23. ISSN 2672-7439

Text 0286225022026103053047.pdf Download (638kB)

Abstract

User Interface (UI) design is a critical component of human– computer interaction in the digital era. However, traditional UI optimization processes still face challenges such as long development cycles, reliance on subjective feedback, and the absence of objective performance metrics. This study proposes an integrated framework that combines real-time eye-tracking analysis with machine learning to optimize UI design in a systematic and data driven manner. A mixed-method approach was applied involving ten participants who interacted with a digital interface prototype while their visual behaviour was recorded using a screen-based eye-tracking system. Quantitative data from gaze coordinates, fixation duration, saccade amplitude, and pupil dilation were analysed using descriptive statistics and supervised machine learning to identify behavioural patterns linked to interface usability. Qualitative feedback was also collected to complement the visual metrics. The results revealed that 72% of total fixation time was concentrated on high information areas such as menus and action buttons, with an average fixation duration of 320 ms, indicating effective visual focus. Pupil dilation increased by 0.4 mm during interaction with new interface elements, reflecting higher cognitive engagement. Furthermore, 85% of users perceived the optimized interface as more intuitive and visually clearer. These findings demonstrate that integrating physiological and behavioural data enhances the objectivity and precision of UI evaluation. The developed framework offers a scalable solution for UI designers to measure and improve interface performance based on empirical evidence rather than subjective assessment. This research contributes to advancing data-driven methodologies in UI/UX optimization and sets a foundation for future integration of multimodal user analytics.

Item Type:	Article
Uncontrolled Keywords:	User interface, Eye tracking, Machine learning, UI optimization, Human-computer interaction
Divisions:	Faculty of Artificial Intelligence and Cyber Security
Depositing User:	Sabariah Ismail
Date Deposited:	13 Jul 2026 04:49
Last Modified:	13 Jul 2026 04:49
URI:	http://eprints.utem.edu.my/id/eprint/29774
Statistic Details:	View Download Statistic

Actions (login required)

View Item