Published Paper Details:

Food is crucial for human life and has been the concern of many healthcare conventions. Nowadays new nutritional assessment and vitamins analysis gear permit more possibilities to assist human beings understand their each day consuming habits, exploring vitamins patterns and hold a wholesome diet. In this paper, we increase a deep model-primarily based totally meals recognition and nutritional assessment gadget to look at and analyze meals gadgets from each day meal pix (e.g., captured by smart phone). Specifically, we recommend a three-step set of rules to understand multi-item (meals) pix by detecting candidate areas and using deep convolutional neural network (CNN) for item classification. The gadget first generates more than one location of proposals on enter pix by applying the Region Proposal Network (RPN) derived from Faster R-CNN model. It then identifies each location of proposals through mapping them into feature maps, and classifies them into unique meals categories, in addition to locating them with inside the original pix. Finally, the gadget will analyze the dietary elements primarily based totally on the popularity results and generate a nutritional assessment document through calculating the amount of calories, fat, carbohydrate and protein.