Application of deep learning for image-based Chinese market food nutrients estimation

Peihua Ma; Chun Pong Lau; Ning Yu; An Li; Jiping Sheng

doi:10.1016/j.foodchem.2021.130994

Application of deep learning for image-based Chinese market food nutrients estimation

Peihua Ma, Chun Pong Lau, Ning Yu, An Li, Jiping Sheng^*

^*Corresponding author for this work

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

51 Citations (Scopus)

Abstract

With commercialization of deep learning (DL) models, daily precision dietary record based on images from smartphones becomes possible. This study took advantage of DL techniques on visual recognition tasks and proposed a suite of big-data-driven DL models regressing from food images to their nutrient estimation. We established and publicized the first food image database from the Chinese market, named ChinaMartFood-109. It contained 10,921 images with 23 nutrient contents, covering 18 main food groups. Inception V3 was optimized using other state-of-the-art deep convolutional neural networks, achieving up to 78 % and 94 % for top-1 and top-5 accuracy, respectively. Besides, this research compared three nutrient estimation algorithms and achieved the best regression coefficient (R²) by normalization + AM compared with arithmetic mean and harmonic mean, validating applicability in practice as well as theory. These encouraging results provide further evidence supporting artificial intelligence in the field of food analysis.

© 2021 Elsevier Ltd. All rights reserved.

Original language	English
Article number	130994
Journal	Food Chemistry
Volume	373
Issue number	Part B
Online published	31 Aug 2021
DOIs	https://doi.org/10.1016/j.foodchem.2021.130994
Publication status	Published - 30 Mar 2022
Externally published	Yes

Funding

We thank the Chinese Scholarship Council for support Peihua Ma’s learning and research. This work is supported by National Key R&D Program of China [No. 2018YFC1603300] and the National Natural Science Foundation of China [Grant No. 71633005]. We also thank Mr. Connie Dai from Boston University and Dr. Ksenia Gerasimova from University of Cambridge for their help in revision of manuscript. We thank the Chinese Scholarship Council for support Peihua Ma's learning and research. This work is supported by National Key R&D Program of China [No. 2018YFC1603300] and the National Natural Science Foundation of China [Grant No. 71633005]. We also thank Mr. Connie Dai from Boston University and Dr. Ksenia Gerasimova from University of Cambridge for their help in revision of manuscript.

Research Keywords

Chinese market food
Convolutional neural network
Deep learning
Food composition
Food image
Food nutrients
Nutrients

Access Output

10.1016/j.foodchem.2021.130994

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@article{20f6956a4c8c41ea9da186ec733e133c,

title = "Application of deep learning for image-based Chinese market food nutrients estimation",

abstract = "With commercialization of deep learning (DL) models, daily precision dietary record based on images from smartphones becomes possible. This study took advantage of DL techniques on visual recognition tasks and proposed a suite of big-data-driven DL models regressing from food images to their nutrient estimation. We established and publicized the first food image database from the Chinese market, named ChinaMartFood-109. It contained 10,921 images with 23 nutrient contents, covering 18 main food groups. Inception V3 was optimized using other state-of-the-art deep convolutional neural networks, achieving up to 78 % and 94 % for top-1 and top-5 accuracy, respectively. Besides, this research compared three nutrient estimation algorithms and achieved the best regression coefficient (R2) by normalization + AM compared with arithmetic mean and harmonic mean, validating applicability in practice as well as theory. These encouraging results provide further evidence supporting artificial intelligence in the field of food analysis.{\textcopyright} 2021 Elsevier Ltd. All rights reserved.",

keywords = "Chinese market food, Convolutional neural network, Deep learning, Food composition, Food image, Food nutrients, Nutrients",

author = "Peihua Ma and Lau, {Chun Pong} and Ning Yu and An Li and Jiping Sheng",

note = "Publisher Copyright: {\textcopyright} 2021 Elsevier Ltd",

year = "2022",

month = mar,

day = "30",

doi = "10.1016/j.foodchem.2021.130994",

language = "English",

volume = "373",

journal = "Food Chemistry",

issn = "0308-8146",

publisher = "Elsevier Ltd.",

number = "Part B",

}

TY - JOUR

T1 - Application of deep learning for image-based Chinese market food nutrients estimation

AU - Ma, Peihua

AU - Lau, Chun Pong

AU - Yu, Ning

AU - Li, An

AU - Sheng, Jiping

PY - 2022/3/30

Y1 - 2022/3/30

N2 - With commercialization of deep learning (DL) models, daily precision dietary record based on images from smartphones becomes possible. This study took advantage of DL techniques on visual recognition tasks and proposed a suite of big-data-driven DL models regressing from food images to their nutrient estimation. We established and publicized the first food image database from the Chinese market, named ChinaMartFood-109. It contained 10,921 images with 23 nutrient contents, covering 18 main food groups. Inception V3 was optimized using other state-of-the-art deep convolutional neural networks, achieving up to 78 % and 94 % for top-1 and top-5 accuracy, respectively. Besides, this research compared three nutrient estimation algorithms and achieved the best regression coefficient (R2) by normalization + AM compared with arithmetic mean and harmonic mean, validating applicability in practice as well as theory. These encouraging results provide further evidence supporting artificial intelligence in the field of food analysis.© 2021 Elsevier Ltd. All rights reserved.

AB - With commercialization of deep learning (DL) models, daily precision dietary record based on images from smartphones becomes possible. This study took advantage of DL techniques on visual recognition tasks and proposed a suite of big-data-driven DL models regressing from food images to their nutrient estimation. We established and publicized the first food image database from the Chinese market, named ChinaMartFood-109. It contained 10,921 images with 23 nutrient contents, covering 18 main food groups. Inception V3 was optimized using other state-of-the-art deep convolutional neural networks, achieving up to 78 % and 94 % for top-1 and top-5 accuracy, respectively. Besides, this research compared three nutrient estimation algorithms and achieved the best regression coefficient (R2) by normalization + AM compared with arithmetic mean and harmonic mean, validating applicability in practice as well as theory. These encouraging results provide further evidence supporting artificial intelligence in the field of food analysis.© 2021 Elsevier Ltd. All rights reserved.

KW - Chinese market food

KW - Convolutional neural network

KW - Deep learning

KW - Food composition

KW - Food image

KW - Food nutrients

KW - Nutrients

UR - http://www.scopus.com/inward/record.url?scp=85118512513&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85118512513&origin=recordpage

U2 - 10.1016/j.foodchem.2021.130994

DO - 10.1016/j.foodchem.2021.130994

M3 - RGC 21 - Publication in refereed journal

C2 - 34731793

AN - SCOPUS:85118512513

SN - 0308-8146

VL - 373

JO - Food Chemistry

JF - Food Chemistry

IS - Part B

M1 - 130994

ER -

Application of deep learning for image-based Chinese market food nutrients estimation

Abstract

Funding

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this