Thesis subject

Recipe Generation and Ingredient Prediction from Food Images

Level:

Research area/discipline: Data Science

Prerequisites: Programming in Python, Big Data, and Machine Learning

Short description:

Food preferences around the World are changing specially in large cities due to factors like increasing income, migration and cultural mixing, appearance of new food fads and life-styles. These dietary changes often imply changes in the quantity and/or types of specific ingredients demanded.
The demand for ingredients, their processing and supplying to consumers often happens in specific locations. Understanding their relevance in recipes used in specific locations can potentially inform design of interventions to improve, among others, nutrition and health, environmental and urban planning. In order to characterize ingredients used in specific locations, innovative tools are being used around the World to generate location-specific datasets.

In this thesis study, the following tasks will be implemented:

  • Systematic Literature Review (SLR) on Recipe Generation from Food Images: This will be performed as part of this thesis study to get information about the state-of-the-art and state-of-the-practice of the techniques on this problem. Available datasets, techniques, features, and challenges & obstacles will be determined with the help of this task.
  • Building Location-Specific Dataset of Street Food using Web Scraping: This task will help to build the dataset for further analysis. At the end of the first task (SLR), it is possible that the other techniques like web scraping can be determined and therefore, instead of only web scraping, the other available approaches can also be used for this task.
  • Recipe Generation and Ingredients Prediction on the Location-Specific Dataset of Street Food: This is the actual work which will be built on the previous tasks. In this task, recipe will be generated and the ingredients will be predicated based on the dataset prepared in previous steps.

Given the focus on both the technological aspects of location-specific ingredients mapping and dietary changes, this thesis is framed as a collaboration between INF and ENP chairs. More specifically, this technology-focused thesis will be led by INF and will count with the involvement of ENP Chair (Raffaele Vignola Raffaele.vignola@wur.nl from the research group on Sustainable Food Systems).

For more information:

cagatay.catal@wur.nl

bedir.tekinerdogan@wur.nl