Food-pics: A database of free images of food for experimental research

The Story:

2014: Jens Blechert, Adrian Meule, Kathrin Ohla and Niko Busch, developed a freely available database of food images and associated normative data for the study of eating and appetite. 
Much research in this field is relying on visual stimulation with food images using methods such as EEG, MEG and fMRI. These neurocognitive methods are sensitive to visual properties such as complexity, brightness, contrast, etc of the images. Likewise, food composition is now understood to modulate neural responses. Therefore, we included data on calories and nutrients in the image database. Also subjective evaluations such as liking (palatability) and wanting (desire to eat) are important parameters that vary with age, body weight, gender and culture. To facilitate informed decisions during selection of food images we provide normative subjective rating data for European and North-american populations of different age ranges, BMI classes and gender.

2018: As some users have expressed interest in an extension of the database to include a more international selection of images, we asked our users to contribute to food-pics. Several colleagues* did so, thanks for your openness and support!   By using their copy right free images and by editing and standardizing them we were able to extend food-pics by further 328 images. These are at your disposition now with associated image properties  like color, size, contrast and complexity (further image parameters will possibly follow in future) under the old licence. Of course again with the request to cite us and and to spread the word.
If you are missing something, or you have suggestions for further images please send them to us (we will analyse image properties for you to allow you to compare them with food-pics images and redistribute them to other users)


More details:

Today, food.pics contains 896 photos displaying food items with simple figure-ground composition, including sweet and tasty foods, high and low calorie, warm and cold dishes, processed and raw foods. It also contains 315 comparable non-food images. Example images for food and non-food categories here

The following attributes are specified for each food picture:
  • Nutrients: mg of protein, fat & carbohydrates, kcal
  • Rating data (palatability, desire to eat, complexity, recognizability, valence, arousal by  almost 2000 participants from German speaking countries and North America)
  • Image properties: R/G/B values, brightness, contrast, complexity (edge detection), luminance, spectral power. Matlab scripts here.

Terms of use:

We provide the data base free of charge on the basis of a license agreement completed by the supervisor/PI/Professor of a work group using the pics. He/she would be responsible for the use of the food-pics in the work group in  keeping with the licence agreement. This is relevant for issues such as storage on shared network spaces, instruction of present and incoming students regarding licence content.

Reference and Citation:

To gain access click here (registration, license, download).

 



* Many thanks to our food-pics_extended contributors:
Loïc P. Heurley, Université Paris Ouest Nanterre La Défense, Nanterre, France
Jang-Han Lee, Chung-Ang University, Seoul, Korea
Gal Sheppes, Tel Aviv University, Tel Aviv, Israel
Loukia Tzavella, Cardiff University, Wales, UK
Olga Pollatos, Universität Ulm, Ulm, Germany
Vaibhav Tyagi, Plymouth University, Devon, UK