International Journal of Multidisciplinary Research and Human Factors

IISSN: 3043-579X

Published: 3^rd November, 2022

DOI:  https://doi.org/10.66419/KPZT9658

 

Performance Metrics of Kanade-Lucas-Tomasi (KLM) filters on Users’ Perception to Graphic Images through
Data Generated from Pupillary Response

 

Fatima ISIAKA

Salihu Aish ABDULKARIM

Abstract

Most research focuses on determining the correlates of pupil dilation and constriction to changes in light intensity in contrast to bright and grey images of a visual interface. It has found that pupil response is highly correlated with high cognitive responses related to problem-solving, akin to the response to strong light intensity. An experiment was conducted with the consent of forty participants, whose pupil responses were recorded in video frames in response to image stimuli. The aim was to generate response data while comparing their pupil responses to bright and coloured images with their greyscale versions. The results indicate that bright and coloured pictures are more appealing and preferable to their grey counterparts, and the pupils tend to dilate in reaction to comprehension. This differs from using changes in light intensity and provides a novel definition of the cognitive response of users to a simple stimulus through pupillary response. The proposed model would aid in generating response data to be used in designing prototypical analytics for assessing and predicting users’ responses to visual images, as well as optimising visual aesthetics and cognitive elucidations.

Keywords: Kanade-Lucas-Tomasi filters, Viola Jones model, Performance metrics, Intelligent analytics, Pupil response, Baseline Response, User Cognition.