Single-View Food Portion Estimation Based on Geometric Models

Abstract

In this paper we present a food portion estimation technique based on a single-view food image used for the estimation of the amount of energy (in kilocalories) consumed at a meal. Unlike previous methods we have developed, the new technique is capable of estimating food portion without manual tuning of parameters. Although single-view 3D scene reconstruction is in general an ill-posed problem, the use of geometric models such as the shape of a container can help to partially recover 3D parameters of food items in the scene. Based on the estimated 3D parameters of each food item and a reference object in the scene, the volume of each food item in the image can be determined. The weight of each food can then be estimated using the density of the food item. We were able to achieve an error of less than 6% for energy estimation of an image of a meal assuming accurate segmentation and food classification.

New mobile methods for dietary assessment: review of image-assisted and image-based dietary assessment methods

“…Ptomey et al (17) analysed whether digital images are a feasible method to improve estimation of energy and macronutrient intake of proxy-assisted 3-d dietary records among adolescents diagnosed with intellectual and developmental disabilities. Twenty adolescents aged between 11 and 18 years with mild (intelligence quotient 50-69) to moderate (intelligence quotient [35][36][37][38][39][40][41][42][43][44][45][46][47][48][49] intellectual and developmental disabilities were given a tablet (iPad 2, Apple, Cupertino, CA, USA) to capture images of all food items consumed over three consecutive days of two weekdays and one weekend day. The participants were instructed to take before and after images by themselves and place a 5 × 5 cm 2 checkered fiduciary marker in the image.…”

“…The portion size estimation relies on the estimated volume of the identified food item. The volume estimation is performed based on three-dimensional reconstruction of the food item from the image as described previously (36)(37)(38)(39) . Images captured over a 24-h day by adolescents (n 15) were used to assess the error of automated determination of food weights compared with the known weights (38) .…”

“…Algorithms used in 2012 (38) Algorithms used in 2015 (39) Crude received the dietary feedback without the weekly text messages and group C (control group) received none.…”

A Mobile Food Record For Integrated Dietary Assessment

“…Each food item in the image is segmented, identified [49], and its volume is estimated [16]. From this information, the energy and nutrient intake can be determined using a food composition table [27].…”

“…These results from the “before” image are sent back to the mFR (step 3) where the user reviews and confirms the food labels (step 4). In step 5, the server receives the information back from the user and uses these results for final image analysis refinement and volume estimation [16]. Step 6 consists of structuring the data generated in the previous steps by forming object descriptions (e.g.…”