Publication: Improving Assistive Software for Color Vision Deficiency through Multiple Model Aggregation

Assistive software that recolors images for individuals with color vision deficiency relies on models of the color differentiation abilities of its intended users. Situation-specific models address the shortcomings of current assumption-based models by using in-situ calibration to capture the color differentiation abilities of a specific user in a specific environment. However, this calibration procedure is time consuming, and when the user is unable to perform it, the assistive software fails to recolor properly. To address this problem, we propose a collection of situation-specific models--Multiple Model Aggregation (MMA)--that can be used to instantly provide the best previously-generated model to the assistive software with no input required from the user. Design challenges for extending MMA to any model-based system are also presented.

Downloads

PDF

Participants

David Flatla
University of Dundee
Carl Gutwin
University of Saskatchewan

Citation

Flatla, D.R., Gutwin, C. 2011. Improving Assistive Software for Color Vision Deficiency through Multiple Model Aggregation. In CHI 2011 Workshop on Dynamic Accessibility, Vancouver, B.C., Canada. CHI 2011 Extended Abstracts.

BibTeX

@inproceedings {215-Flatla_CHI2011_WS24,
author= {David Flatla and Carl Gutwin},
title= {Improving Assistive Software for Color Vision Deficiency through Multiple Model Aggregation},
booktitle= {CHI 2011 Workshop on Dynamic Accessibility},
year= {2011},
address= {Vancouver, B.C., Canada},
note= {CHI 2011 Extended Abstracts}
}