Abstract

Well characterized databases used for experimental purposes with extensive meta-data are essential for conducting meaningful and comparable studies. The Food-pics_extended database is one example for a widely used food stimulus database. Indeed, meta-data on low level and high level image characteristics is broad, yet fiber ratings are not included, limiting its use in diet-related studies. Therefore, we developed fiber ratings per item, based on mean values of four non-expert raters. Ratings show good reliability and meaningful ranges per food type. The newly provided fiber ratings enrich the already valuable database and extend it by an important nutrient value for human and planetary health.