Food Dataset for Food Recognition:
Food image recognition is one of the promising applications of object recognition technology, since it will help estimate food calories and analyze people’s eating habits for healthcare.
- http://academictorrents.com/details/fbc7a9f9a10be134a1738ba947efa1814ed3ce9b ImageNet 2014 – good quality of labels, many examples of dishes (1.2k+ for fish and chips) 50GB universal dataset
- https://www.vision.ee.ethz.ch/datasets_extra/food-101/ “Food-101 (2014) – contains some errors in labels (Swiss)” 5GB
- http://foodcam.mobi/dataset256.html (Japanese food dataset) “UEC FOOD-100 UEC FOOD-256” (non-commercial research purpose)
- http://www.site.uottawa.ca/~shervin/food/ (Canada / Ind) FooDD: Food Detection Dataset)
Food Dataset for Voice Applications:
- https://datahub.io/dataset/fooddb FoodDB (1: ingredient / dish name, 2: preparation mode) 9173 entries (Voice search)
Recipes and ingredients Datasets:
- https://datahub.io/dataset/recipe-dataset – recipes and ingredients – The source data, in CSV, was taken from the paper “Flavor network and the principles of food pairing”. It is a collection of recipes from http://allrecipes.com, http://epicurious.com and http://menupan.com. Each recipe has an associated cuisine value (one of 11 world cuisines), and a flattened list of ingredients.
Other Food Datasets:
- https://datahub.io/dataset/foodpedia – FOODpedia – Linked Data Dataset about Food Products and Ingredients
- https://datahub.io/dataset/health-nutrition-products – heath and nutrition Health & Nutrition Products
- https://datahub.io/dataset/foodista – foods and recipes – Foodista is a community edited recipe wiki, published under a Creative Common Attribution license. The wiki contains information on foods, tools, techniques, and recipes. Copy: https://archive.org/details/kasabi
- https://datahub.io/dataset/open-food-facts – copy of: http://world.openfoodfacts.org/ – Image dataset of branded food products
Food Recognition: Products, Companies, API:
- http://blog.clarifai.com/what-food-is-this-clarifais-food-recognition-technology-can-tell-you/ – ClarifAI
- https://www.cs.ubc.ca/~murphyk/Papers/im2calories_iccv15.pdf – Im2Calories (Google version) – Segmentations on some images from the Food201-Segmented test set. App: https://www.facebook.com/Im2Calories-Google-app-1621161728130123/ Paper: “Im2Calories: towards an automated mobile vision food diary” https://www.cs.ubc.ca/~murphyk/Papers/im2calories_iccv15.pdf (PDF). Youtube demo: https://www.youtube.com/watch?v=7o_aM3zMN8Q
Demo #2: https://www.youtube.com/watch?v=KbgG_82e1UI“ - https://blogs.nvidia.com/blog/2016/09/30/deep-learning-automatic-calorie-counter/ LoseIt! SnapIt! Youtube Demo: https://www.youtube.com/watch?v=WuZ6gHoX20c – Accuracy rate is about 87 percent for foods commonly entered by its users. That surpassed others tested using the standard measure in the Food-101 dataset.
Some Papers on Food Recognition:
- http://cs229.stanford.edu/proj2015/233_report.pdf – Cuisine Classification and Recipe Generation
- http://ubicomp.org/ubicomp2014/proceedings/ubicomp_adjunct/workshops/CEA/p589-kawano.pdf – Food Image Recognition with Deep Convolutional Features
- https://www.researchgate.net/publication/279850763_FooDD_Food_Detection_Dataset_for_Calorie_Measurement_Using_Food_Images – Food Image Recognition with Deep Convolutional Features