Visualization of Texture Expressions for Recipes by Using Reviews

Nowadays, there are lots of recipe sites on the net, work to inspire their users with visual and step-by-step guided cooking instructions. In these sites, recipe retrieval mechanisms are provided so that users can find recipes for dishes they want to cook by specifying their ingredients and nutrition facts or using keywords. In the circumstance as mentioned above, this paper focuses on the further question of the user: "What would be the result when the user cooks with this recipe?" A user's cooking review of the recipe says what happened when the reviewer tried the recipe. Not a few recipes have more than a thousand reviews. In this paper, we discuss recipe representation that extracts and aggregates texture terms from a large number of reviews of the recipe. This research aims to propose a visualization method that allows users to imagine the result of their cooking by following the recipe intuitively. Finally, we will show its visualization using a tag cloud.

I. Introduction

Eating texture refers to the skin sensations in the mouth, including through teeth and the tongue, that one feels when eating or drinking food. According to ISO11036, texture is "All the mechanical, geometrical and surface attributes of a product perceptible by means of mechanical, tactile and, where appropriate, visual and auditory receptors." Even in online recipes, texture terms are used for titles, directions, and descriptions. There are many words to express textures in each language. By using texture to describe a recipe, it is possible to easily convey the result to the users in a limited number of characters. The rapid development of information retrieval technology has been bringing a variety of new services to the Internet. Internet recipe sites also provide some useful recipe retrieval features. Users can find recipes on what they want to cook by specifying the ingredients or nutrition facts and by using keywords in the titles, descriptions, and directions of the recipes. Recipe websites have become a popular way to exchange and/or share information about cooking. There are two types of users: recipe authors and readers. Recipe authors are servers; readers are clients. Both of them can exchange and share information about various recipes via reviews and comments. These reviews are written about their experiences when they attempted the recipe. We investigate the elaborate use of the reviews to help explore the user question: "What will be the result when I cook with this recipe?" Our research aims to provide a visualization method to show the result of the recipe. We suggest a framework to show our users who are trying a recipe for the first time what happened to users who have already cooked with it. First, we discuss a method for evaluating a recipe by extracting texture terms from its reviews. Then we show how we obtain a texture expression for the recipe. The visualization of a recipe presents the final result. This paper has 5 sections. Related works, which clarify the position of the study, are introduced in the second section. The third section is concerned with the new method of recipe evaluation conducted in this research. We discuss the recipe presentation method in the fourth section and end our paper with the conclusion in the fifth section.