Models are built using data. Most successful search, social media, and recommendation systems are built using personal models to provide people the right information, at the right time, in the right context, usually even before a user articulates his need. Food recommendation systems need to be built using the same approach. A personal food model is essential for recommending the right food item at the right time. It is also essential to predict the effect of food so the right suggestions can be made to avoid unpleasant situations. We need to build such models using food logs collected for the person. Many applications for foodlogging are being developed based on detecting the dish or item being consumed and finding nutritional elements based on ingredients in these items. Detecting items and the volumes consumed requires a multimodal platform and nutritional data sources for items prepared using specific ingredients and recipes.
– Ruihan Xu, Luis Herranz, Shuqiang Jiang, Shuang Wang, Xinhang Song, Ramesh Jain, Geolocalized Modeling for Dish Recognition. IEEE Transactions on Multimedia 17(8): 1187-1199 (2015).
– Nitish Nag, Aditya Narendra Rao, Akash Kulhalli, Kushal Samir Mehta, Nishant Bhattacharya, Pratul Ramkumar, Aditya Bharadwaj, Dinkar Sitaram, Ramesh C. Jain, Flavour Enhanced Food Recommendation. MADiMa @ ACM Multimedia 2019: 60-66.
– Nitish Nag, Vaibhav Pandey, Ramesh C. Jain, Health Multimedia: Lifestyle Recommendations Based on Diverse Observations. ICMR 2017: 99-106.