Food weight and nutrient content identification method based on image recognition

Abstract

The invention relates to a food weight and nutrient content identification method based on image recognition. The method comprises the steps that a target food picture is acquired; point acquiring andmarking are carried out on the acquired food picture; a selected picture path and the data of point acquiring and marking are calculated, and each value of point acquiring and marking is labeled; thetarget food picture is acquired according to the picture path, and color extraction and color modeling are carried out on a point acquiring place according to the data and label of point acquiring and marking; a target food is scanned, and color matching is carried out to complete the classification of various foods; scanning and matching are carried out on a terminal; a recognition algorithm iscalled to calculate the area of the target food; and through the acquired area data and the acquired density and nutrients of the target food, the weight of the target food and the nutrients containedin the target food are calculated to acquire the weight of the target food and the nutrients contained in the target food. The method provided by the invention has the advantages of simple and quickoperation and high recognition accuracy.

Description

technical field [0001] The invention relates to the technical field of computer applications, in particular to an identification method for food weight and nutrient content based on image recognition. Background technique [0002] At present, there are few algorithms or software on the market that can do food image recognition to obtain the weight and nutrient content of each food. Some recognition algorithms or software can only recognize food weight with an accuracy of about 50%. When image recognition obtains food weight and nutrients, the error may reach more than 40%. The accuracy of food weight recognition can only reach about 50%, which leads to: a. It is completely unusable for users who have informed requirements on the weight and nutrients of each meal; b. If there are one or more types of food on the plate When the colors are similar, it is also easy to cause recognition errors in the algorithm or to regard similar colors as a kind of food. The reason for the si...

AI Technical Summary

Problems solved by technology

The accuracy of food weight recognition can only reach about 50%, which leads to: a. It is completely unusable for users who have informed requirements on the weight and nutrients of each meal; b. If there are one or more types of food on the plate When the colors are similar, it is also easy to cause recognition errors in the algorithm or to regard similar colors as a kind of food

The reason for the situation a is that their algorithms or software focus on calculating the volume of each food in the food image, which requires a two-view 3D reconstruction method to achieve, which is very complicated and difficult

The reason for case b is that their algorithm does not model and segment each food color

Images