Combining deep residual network features with supervised machine learning algorithms to classify diverse food image datasets

Patrick McAllister, Huiru Zheng, RR Bond, Anne Moorhead

Research output: Contribution to journalArticle

12 Citations (Scopus)

Abstract

Obesity is increasing worldwide and can cause many chronic conditions such as type-2 diabetes, heart disease, sleep apnea, and some cancers. Monitoring dietary intake through food logging is a key method to maintain a healthy lifestyle to prevent and manage obesity. Computer vision methods have been applied to food logging to automate image classification for monitoring dietary intake. In this work we applied pretrained ResNet-152 and GoogleNet convolutional neural networks (CNNs) to extract features from food image datasets; Food 5K, Food-11, RawFooT-DB, and Food-101. Deep features were extracted from CNNs and used to train machine learning classifiers including artificial neural network(ANN), support vector machine(SVM), Random Forest, fully connected Neural Networks, and Naive Bayes. Results show that using ResNet-152 deep features with SVM with RBF kernel can accurately detect food items with 99.4% accuracy using Food-5K food image dataset. Trained with ResNet-152 features, ANN can achieve 91.34%, 99.28% when applied to Food-11 and RawFooT-DB food image datasets respectively and SVM with RBF kernel can achieve 64.98% with Food-101 image dataset. From this research it is clear that using deep CNN features can be used efficiently for diverse food item image classification. The work presented in this research shows that pretrained ResNet-152 features provide sufficient generalisation power when applied to a range of food image classification tasks.
LanguageEnglish
JournalComputers in Biology and Medicine
Volume1
Early online date17 Feb 2018
DOIs
Publication statusE-pub ahead of print - 17 Feb 2018

Fingerprint

Learning algorithms
Learning systems
Food
Neural networks
Image classification
Support vector machines
Datasets
Supervised Machine Learning
Obesity
Sleep Apnea Syndromes
Monitoring
Research
Medical problems
Type 2 Diabetes Mellitus
Computer vision
Heart Diseases
Eating
Classifiers

Keywords

  • Obesity
  • Food logging
  • Deep learning
  • Convolutional neural networks
  • Feature extraction

Cite this

@article{15f60f269d874fa1a104d6d2dc359386,
title = "Combining deep residual network features with supervised machine learning algorithms to classify diverse food image datasets",
abstract = "Obesity is increasing worldwide and can cause many chronic conditions such as type-2 diabetes, heart disease, sleep apnea, and some cancers. Monitoring dietary intake through food logging is a key method to maintain a healthy lifestyle to prevent and manage obesity. Computer vision methods have been applied to food logging to automate image classification for monitoring dietary intake. In this work we applied pretrained ResNet-152 and GoogleNet convolutional neural networks (CNNs) to extract features from food image datasets; Food 5K, Food-11, RawFooT-DB, and Food-101. Deep features were extracted from CNNs and used to train machine learning classifiers including artificial neural network(ANN), support vector machine(SVM), Random Forest, fully connected Neural Networks, and Naive Bayes. Results show that using ResNet-152 deep features with SVM with RBF kernel can accurately detect food items with 99.4{\%} accuracy using Food-5K food image dataset. Trained with ResNet-152 features, ANN can achieve 91.34{\%}, 99.28{\%} when applied to Food-11 and RawFooT-DB food image datasets respectively and SVM with RBF kernel can achieve 64.98{\%} with Food-101 image dataset. From this research it is clear that using deep CNN features can be used efficiently for diverse food item image classification. The work presented in this research shows that pretrained ResNet-152 features provide sufficient generalisation power when applied to a range of food image classification tasks.",
keywords = "Obesity, Food logging, Deep learning, Convolutional neural networks, Feature extraction",
author = "Patrick McAllister and Huiru Zheng and RR Bond and Anne Moorhead",
year = "2018",
month = "2",
day = "17",
doi = "10.1016/j.compbiomed.2018.02.008",
language = "English",
volume = "1",
journal = "Computers in Biology and Medicine",
issn = "0010-4825",
publisher = "Elsevier",

}

TY - JOUR

T1 - Combining deep residual network features with supervised machine learning algorithms to classify diverse food image datasets

AU - McAllister, Patrick

AU - Zheng, Huiru

AU - Bond, RR

AU - Moorhead, Anne

PY - 2018/2/17

Y1 - 2018/2/17

N2 - Obesity is increasing worldwide and can cause many chronic conditions such as type-2 diabetes, heart disease, sleep apnea, and some cancers. Monitoring dietary intake through food logging is a key method to maintain a healthy lifestyle to prevent and manage obesity. Computer vision methods have been applied to food logging to automate image classification for monitoring dietary intake. In this work we applied pretrained ResNet-152 and GoogleNet convolutional neural networks (CNNs) to extract features from food image datasets; Food 5K, Food-11, RawFooT-DB, and Food-101. Deep features were extracted from CNNs and used to train machine learning classifiers including artificial neural network(ANN), support vector machine(SVM), Random Forest, fully connected Neural Networks, and Naive Bayes. Results show that using ResNet-152 deep features with SVM with RBF kernel can accurately detect food items with 99.4% accuracy using Food-5K food image dataset. Trained with ResNet-152 features, ANN can achieve 91.34%, 99.28% when applied to Food-11 and RawFooT-DB food image datasets respectively and SVM with RBF kernel can achieve 64.98% with Food-101 image dataset. From this research it is clear that using deep CNN features can be used efficiently for diverse food item image classification. The work presented in this research shows that pretrained ResNet-152 features provide sufficient generalisation power when applied to a range of food image classification tasks.

AB - Obesity is increasing worldwide and can cause many chronic conditions such as type-2 diabetes, heart disease, sleep apnea, and some cancers. Monitoring dietary intake through food logging is a key method to maintain a healthy lifestyle to prevent and manage obesity. Computer vision methods have been applied to food logging to automate image classification for monitoring dietary intake. In this work we applied pretrained ResNet-152 and GoogleNet convolutional neural networks (CNNs) to extract features from food image datasets; Food 5K, Food-11, RawFooT-DB, and Food-101. Deep features were extracted from CNNs and used to train machine learning classifiers including artificial neural network(ANN), support vector machine(SVM), Random Forest, fully connected Neural Networks, and Naive Bayes. Results show that using ResNet-152 deep features with SVM with RBF kernel can accurately detect food items with 99.4% accuracy using Food-5K food image dataset. Trained with ResNet-152 features, ANN can achieve 91.34%, 99.28% when applied to Food-11 and RawFooT-DB food image datasets respectively and SVM with RBF kernel can achieve 64.98% with Food-101 image dataset. From this research it is clear that using deep CNN features can be used efficiently for diverse food item image classification. The work presented in this research shows that pretrained ResNet-152 features provide sufficient generalisation power when applied to a range of food image classification tasks.

KW - Obesity

KW - Food logging

KW - Deep learning

KW - Convolutional neural networks

KW - Feature extraction

U2 - 10.1016/j.compbiomed.2018.02.008

DO - 10.1016/j.compbiomed.2018.02.008

M3 - Article

VL - 1

JO - Computers in Biology and Medicine

T2 - Computers in Biology and Medicine

JF - Computers in Biology and Medicine

SN - 0010-4825

ER -