eLife assessment
In this important study, the authors manually assessed randomly selected images published in eLife between 2012 and 2022 to determine whether they were accessible for readers with deuteranopia, the most common form of color vision deficiency. They then developed an automated tool designed to classify figures and images as either "friendly" or "unfriendly" for people with deuteranopia. Such a tool could be used by journals or researchers to monitor the accessibility of figures and images, and the evidence for its utility was solid: it performed well for eLife articles, but performance was weaker for a broader dataset of PubMed articles, which were not included in the training data. The authors also provide code that readers can download and run to test their own images, and this may be of most use for testing the tool, as there are already several free, user-friendly recoloring programs that allow users to see how images would look to a person with different forms of color vision deficiency. Automated classifications are of most use for assessing many images, when the user does not have the time or resources to assess each image individually.