Towards Eating Habits Discovery in Egocentric Photo-Streams

Alina Matei,Andreea Glavan,Petia Radeva,Estefania Talavera

doi:10.1109/access.2021.3053175

Abstract

Eating habits are learned throughout the early stages of our lives. However, it is not easy to be aware of how our food-related routine affects our healthy living. In this work, we address the unsupervised discovery of nutritional habits from egocentric photo-streams. We build a food-related behavioral pattern discovery model, which discloses nutritional routines from the activities performed throughout the days. To do so, we rely on Dynamic-Time-Warping for the evaluation of similarity among the collected days. Within this framework, we present a simple, but robust and fast novel classification pipeline that outperforms the state-of-the-art on food-related image classification with a weighted accuracy and F-score of 70% and 63%, respectively. Later, we identify days composed of nutritional activities that do not describe the habits of the person as anomalies in the daily life of the user with the Isolation Forest method. Furthermore, we show an application for the identification of food-related scenes when the camera wearer eats in isolation. Results have shown the good performance of the proposed model and its relevance to visualize the nutritional habits of individuals.

Highlights

Nutrition has a significant influence on everyone’s daily routine
We study the pattern of occurrence of the food-related images to find nutritional habits
We depict the model for nutritional habits discovery, which relies on the assigned labels to food-related images

Summary

Introduction

Studies have shown that American people spend on average 2.5h a day eating or drinking, out of which 78 minutes eating or drinking while doing other primary activities, such as driving, preparing meals or working [1]. What people eat has been regarded so far as the main factor impacting people’s food behavior. Recent studies have shown that how and where people eat play an important role [2]. Our hypothesis is that by gathering insight into the context of food-related activities, people can improve their eating habits with the view of leading a healthier lifestyle. Eating habits have a direct impact on one’s health: diseases like diabetes, obesity, cardiovascular conditions, cancer and even mental illnesses are closely related to what people eat [3]–[5]. Research has shown that loneliness has links to eating disorders [6], while television advertisements that promote calorie-dense foods and snacks have been shown to trigger mindless eating or snacking [7]

Objectives

Methods

Results

Conclusion