Web Project

Investigating the Homogenization of

Web Design Using Computer Vision

We’re looking for participants for an ongoing study! If you or someone you know has been involved in web design since the mid-2000’s, please find out more here!

HCI is interested in how the everyday interfaces we encounter shape our perception of society. For example, search engines help us more easily navigate the web, but their centralization gives them the ability to perpetuate and normalize harmful societal biases. Developing tools to identify centralized control over our interfaces is an important first step towards understanding the potential harms. In this paper, we investigate the growing concern that visual design of the World Wide Web has homogenized over the past decade by developing automatic techniques to characterize and quantify this possible homogenization. In particular, we apply two types of computer vision techniques, deep learning with convolutional neural networks and hand-engineered color and layout features, each with complementary strengths and weaknesses. Our results show that designs have become more similar since 2007, especially for page layouts, where the average distance between sites decreased by over 30%. We investigate potential causes of this homogenization including overlap in source code and web libraries, color scheme standardization, and support for mobile devices. We hope our results motivate further discussion of the factors which influence designers and the implications of these factors on the future trajectory of the web.

An example of our hand-engineered visual representations: (a) a sample website from the Alexa dataset, (b) area plot and circular hue histogram showing the website’s color distribution, and (c) nodes of the XY-tree decomposition, visualized as red boxes overlaid on the raw image.
: Average pairwise similarity between the Russell-100 websites, plotted as a function of time, for color, layout, and CNN features. Shaded regions show 95% confidence intervals. To allow comparison across the three metrics (which have different dimensions), each plot was normalized to 0 mean and unit standard deviation. In each plot, higher values are more homogeneous. All three metrics show increases in the period 2008-201

Arguably, the homogenization of design may signal that a few corporations have gained influence over what constitutes proper design on web.

This project is ongoing research! We are preparing our work for publication and plan to update this page with code and data later on.

The IU Computer Vision Lab's projects and activities have been funded, in part, by grants and contracts from the Air Force Office of Scientific Research (AFOSR), the Defense Threat Reduction Agency (DTRA), Dzyne Technologies, EgoVid, Inc., ETRI, Facebook, Google, Grant Thornton LLP, IARPA, the Indiana Innovation Institute (IN3), the IU Data to Insight Center, the IU Office of the Vice Provost for Research through an Emerging Areas of Research grant, the IU Social Sciences Research Commons, the Lilly Endowment, NASA, National Science Foundation (IIS-1253549, CNS-1834899, CNS-1408730, BCS-1842817, CNS-1744748, IIS-1257141, IIS-1852294), NVidia, ObjectVideo, Office of Naval Research (ONR), Pixm, Inc., and the U.S. Navy. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the U.S. Government, or any sponsor.