Date of Award


Degree Name

Doctor of Philosophy


Industrial and Entrepreneurial Engineering and Engineering Management

First Advisor

Dr. Steven Butt

Second Advisor

Dr. Tycho Fredericks

Third Advisor

Dr.Lee Wells

Fourth Advisor

Dr. Ikhlas Abdel-Qader


Pressure mapping, data analytics, image registration, machine learning, image processing, seating research


The technological advancements in sensors, monitoring systems, and tracking devices are changing how we study our environment; big data sets are becoming more and more prevalent due to the increase of information gathered with ease. One system benefiting from these technological improvements is pressure mapping technology, an easy-to-use and cost-effective solution for assessing contact pressure distributions.

Pressure mapping systems generally produce data sets of very large volume, especially when used for continuous tracking and monitoring, and are widely used for research in fields of ergonomics, sports, industries, and health disciplines. Pressure mapping systems are particularly important in the study of human-chair seating interactions. Researchers have widely used pressure mapping systems to study these interactions and their relationship with comfort/discomfort across different conditions. The analysis of seating pressure maps usually consists in evaluating descriptive pressure measures and using visual feedback for assessing pressure distributions. Unfortunately, current analytical techniques do not provide clear insights about pressure distribution patterns nor spatial relationships within seating pressure maps; these are needed to further understand human-chair interactions. The need for additional pressure distribution measures, along with quantitative techniques for assessing and comparing pressure maps, have also been emphasized in literature.

This work studies the applications of machine learning, spatial data analytics, digital image processing, and optimal image registration as new techniques for pressure mapping analysis, with the objective of implementing these techniques to pre-process, analyze, and compare seating pressure map images. The results of this study demonstrate the practicality and effectiveness of using these techniques for (1) removing extrinsic pressure artifacts (outliers) by using density-based spatial clustering, (2) measuring distribution patterns and spatial relationships by using spatial autocorrelation and statistical features of images, and (3) aligning and comparing pressure map by using image registration and similarity/dissimilarity coefficients.

The use of DBSCAN and DENCLUE clustering algorithms were found to be suitable for identifying and eliminating extrinsic pressure artifacts (outliers), with obtained overall accuracies over ninety-nine percent. Moran’s I spatial autocorrelation measure, and image statistical features of Skewness, Correlation (GLSD), Gradient Contrast/Mean (GLD), Gradient Second Moment (GLD), and Homogeneity (GLSD) were found to be appropriate for measuring unique aspects of pressure distributions within pressure maps. Image registration based on the minimization of the Mean Square Error (MSE) was also suitable for aligning pressure map images, with similarity and dissimilarity coefficients of Pearson Correlation Coefficient, Minimum Ratio, 𝐿1 Norm, and Intensity Ratio Variance being particularly unique when comparing aligned pressure maps.

These methodologies can help future seating research by providing additional analytical tools for a better understanding of user-chair interactions and their relationships with sitting comfort/discomfort, in both static and dynamic sitting environments. While findings in this study are in the context of task seating (i.e. mousing and typing), these techniques can also be tailored and employed in other seating research applications (e.g., automobile seating, aircraft pilot seats, and paraplegic seating), non-seating pressure map research (e.g., gait analysis, industrial applications, and sports fields), or research studies using spatially related three-dimensional datasets (e.g., surface topography, contour data, and heat maps).

Access Setting

Dissertation-Open Access