Line Recognition for Generating Accessible Line Plots

Existing technologies identify individual lines in a line plot by matching the symbols in the legend. Identifying each symbol is crucial to create accessible graphics for people who are blind so that they provide the best resemblance possible to the original plot. This resemblance provides better communication between sighted and vision impaired users. In our work we use an existing network architecture and improve it in a way that it will get more semantic information out of the lines and their symbology. We present these line graphs in the GraVVITAS system, an approach for providing dynamic and refreshable accessible graphics, to be used by people who are blind.