Please note: This PhD seminar will take place online.
Andreea Pocol, PhD candidate
David R. Cheriton School of Computer Science
Supervisors: Professors Lesley Istead, Craig S. Kaplan
How does a machine translate the visual complexity of a digital image into a harmonious and meaningful colour palette? This seminar explores the intersection of computer science and colour psychology to answer that question. It will begin with an attempt to define ‘colour’ and ‘colour palette.’ It will then delve into the long-standing “Universalist” vs. “Culturalist” debate in colour psychology. The talk will synthesize nearly a century of research, ranging from the early studies of Odbert (1942) and Wexner (1954) to modern neuro-aesthetic inquiries by Mohr (2022) and Damiano (2023). The core of the seminar provides an overview of seven distinct computational methods for colour palette extraction, evaluating their strengths and limitations: Colour Clustering (including K-Means, Median Cut, Mean Shift, Octree Quantization, DBSCAN), Histogram Thresholding, and Image Segmentation. This talk offers a comprehensive framework for understanding how we condense coloured media into functional, emotive palettes, and the impact thereof on viewer affect.