Graphical recipe automation

Graphical recipes are a way of representing cooking instructions and we wanted to make the process of making them automatic. The software would be able to read in a recipe text, ingredient list and instructions, and output a graphical recipe. To extract information out of written text Natural Language Processing was used. We ran into some trouble with part of speech tagging the text and in the end trained our own tagger with an accuracy of 97%. The software does not generate perfect graphical recipes and some changes are needed until it is complete. The automatic conversion process is very quick and its purpose is to help people in making graphical recipes. Using Natural Language Processing we were able to generate graphical recipes that were on average 70-80\% accurate