Accessibility: Automated detection, indication and fixing of content for persons with color blindness. (04-Nov-2009)

Accessibility: Automated detection, indication and fixing of content for persons with color blindness.

Accessible documents with color are required to provide contrast that is sufficient and can be correctly understood by a person who is color blind. Traditionally this test is done by a human using a tool that forces the color content to a grey scale equivalent. The human converts the content to the grey scale version, views the document and determines if any content is unreadable in the grey format. Although this is an effective test, it is a manual test and must be done by a human. This takes both time, some (albeit minimal) training and some domain knowledge.

What is needed is an automated way to detect color blindness problems in images and documents. Further it would be valuable to provide automatic indication of problem areas, and further valuable to provide automatic correction for such problems. The publication proposes an automated method for detecting color blindness related problems in images and documents. The publication also provides for automatic correction of most problems. The publication further provides for automatic markup of images or documents to indicate problem areas. Indicated problem areas can be used to fix up known problem areas, in which case the content can be resubmitted for analysis.

The idea uses a set of techniques to discover and identify text in images, and portions of diagrams which will be difficult for persons with color blindness to detect and understand. The specific steps are outlined in section 3. They generally involve analysis of text, object and images by looking for edge conditions, converting to a value scale and testing for sufficient contrast. This idea is invoked on a running application. It may be invoked as needed by a tester or with each displayed dialog. In the past, automated checkers have been unable to test for color blindness problems. However, this idea can be used in concert with other automation testing tools. Failing dialogs are recorded, and failing words or sections of the dialog are logged. In one embodiment a screen shot of the screen is saved with the failing content identified by a surrounding box. Logging also provides suggestions for changes that would bring the application into compliance.

In one embodiment, this idea is embedded in a document or presentation creation environment. In this embodiment, the testing also includes the option to fix several problems (as outlined in section 3). In another embodiment, the automated checking and testing process is used as a browser plug-in preprocessor. In this embodiment, the user may be a tester, or may be a color-blind user. The system identifies problems so the color-blind use can be aware that more information is available than they are seeing. The user may select an auto-correct mode which can...