Aratraductor: text to pictogram translation using natural language processing techniques

Nowadays, there is a significant amount of the population that have problems when communicating through written text. This means that many people are in high risk of social exclusion as they find themselves without access to key forms of communication such as writing and reading. Besides, this form of communication has become more relevant in relation to new technologies. This paper focuses on providing a tool for translating Spanish text into pictograms that aims to facilitate access to written information for people who cannot use natural language due to physical or mental disabilities. In order to improve the obtained translations we have used Natural Language Processing (NLP) techniques. The evaluation of the system effectively confirms that obtained translations are not only accurate but closer to human-made translations than the ones obtained by a similar tool.