Color Assessment and Transfer for Web Pages

TL;DR

This paper develops a machine learning-based approach to assess and transfer color themes in Web pages, improving visual aesthetics and user interaction by leveraging color assessment models and color transfer techniques.

Contribution

It introduces the first comprehensive method for Web page color assessment and transfer, combining Web mining, computer graphics, and machine learning.

Findings

Color assessment models are effective in evaluating Web page color schemes.

The models successfully guide color transfer to enhance Web page aesthetics.

Experimental results demonstrate the practical utility of the proposed approach.

Abstract

Colors play a particularly important role in both designing and accessing Web pages. A well-designed color scheme improves Web pages' visual aesthetic and facilitates user interactions. As far as we know, existing color assessment studies focus on images; studies on color assessment and editing for Web pages are rare. This paper investigates color assessment for Web pages based on existing online color theme-rating data sets and applies this assessment to Web color edit. This study consists of three parts. First, we study the extraction of a Web page's color theme. Second, we construct color assessment models that score the color compatibility of a Web page by leveraging machine learning techniques. Third, we incorporate the learned color assessment model into a new application, namely, color transfer for Web pages. Our study combines techniques from computer graphics, Web mining, computer vision, and machine learning. Experimental results suggest that our constructed color assessment models are effective, and useful in the color transfer for Web pages, which has received little attention in both Web mining and computer graphics communities.