Mid‐level deep Food Part mining for food image recognition
There has been a growing interest in food image recognition for a wide range of applications. Among existing methods, mid‐level image part‐based approaches show promising performances due to their suitability for modelling deformable food parts (FPs). However, the achievable accuracy is limited by t...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Wiley
2018-04-01
|
Series: | IET Computer Vision |
Subjects: | |
Online Access: | https://doi.org/10.1049/iet-cvi.2016.0335 |
_version_ | 1797684930214363136 |
---|---|
author | Jiannan Zheng Liang Zou Z. Jane Wang |
author_facet | Jiannan Zheng Liang Zou Z. Jane Wang |
author_sort | Jiannan Zheng |
collection | DOAJ |
description | There has been a growing interest in food image recognition for a wide range of applications. Among existing methods, mid‐level image part‐based approaches show promising performances due to their suitability for modelling deformable food parts (FPs). However, the achievable accuracy is limited by the FP representations based on low‐level features. Benefiting from the capacity to learn powerful features with labelled data, deep learning approaches achieved state‐of‐the‐art performances in several food image recognition problems. Both mid‐level‐based approaches and deep convolutional neural networks (DCNNs) approaches clearly have their respective advantages, but perhaps most importantly these two approaches can be considered complementary. As such, the authors propose a novel framework to better utilise DCNN features for food images by jointly exploring the advantages of both the mid‐level‐based approaches and the DCNN approaches. Furthermore, they tackle the challenge of training a DCNN model with the unlabelled mid‐level parts data. They accomplish this by designing a clustering‐based FP label mining scheme to generate part‐level labels from unlabelled data. They test on three benchmark food image datasets, and the numerical results demonstrate that the proposed approach achieves competitive performance when compared with existing food image recognition approaches. |
first_indexed | 2024-03-12T00:36:56Z |
format | Article |
id | doaj.art-250c7adaee7540a58f85ec2797e2857b |
institution | Directory Open Access Journal |
issn | 1751-9632 1751-9640 |
language | English |
last_indexed | 2024-03-12T00:36:56Z |
publishDate | 2018-04-01 |
publisher | Wiley |
record_format | Article |
series | IET Computer Vision |
spelling | doaj.art-250c7adaee7540a58f85ec2797e2857b2023-09-15T09:32:17ZengWileyIET Computer Vision1751-96321751-96402018-04-0112329830410.1049/iet-cvi.2016.0335Mid‐level deep Food Part mining for food image recognitionJiannan Zheng0Liang Zou1Z. Jane Wang2Department of Electrical and Computer EngineeringUniversity of British Columbia5500‐2332 Main MallVancouverCanadaDepartment of Electrical and Computer EngineeringUniversity of British Columbia5500‐2332 Main MallVancouverCanadaDepartment of Electrical and Computer EngineeringUniversity of British Columbia5500‐2332 Main MallVancouverCanadaThere has been a growing interest in food image recognition for a wide range of applications. Among existing methods, mid‐level image part‐based approaches show promising performances due to their suitability for modelling deformable food parts (FPs). However, the achievable accuracy is limited by the FP representations based on low‐level features. Benefiting from the capacity to learn powerful features with labelled data, deep learning approaches achieved state‐of‐the‐art performances in several food image recognition problems. Both mid‐level‐based approaches and deep convolutional neural networks (DCNNs) approaches clearly have their respective advantages, but perhaps most importantly these two approaches can be considered complementary. As such, the authors propose a novel framework to better utilise DCNN features for food images by jointly exploring the advantages of both the mid‐level‐based approaches and the DCNN approaches. Furthermore, they tackle the challenge of training a DCNN model with the unlabelled mid‐level parts data. They accomplish this by designing a clustering‐based FP label mining scheme to generate part‐level labels from unlabelled data. They test on three benchmark food image datasets, and the numerical results demonstrate that the proposed approach achieves competitive performance when compared with existing food image recognition approaches.https://doi.org/10.1049/iet-cvi.2016.0335midlevel deep FP miningfood image recognitionmidlevel image part-based approachesdeformable food parts modellingFP representationsdeep learning approaches |
spellingShingle | Jiannan Zheng Liang Zou Z. Jane Wang Mid‐level deep Food Part mining for food image recognition IET Computer Vision midlevel deep FP mining food image recognition midlevel image part-based approaches deformable food parts modelling FP representations deep learning approaches |
title | Mid‐level deep Food Part mining for food image recognition |
title_full | Mid‐level deep Food Part mining for food image recognition |
title_fullStr | Mid‐level deep Food Part mining for food image recognition |
title_full_unstemmed | Mid‐level deep Food Part mining for food image recognition |
title_short | Mid‐level deep Food Part mining for food image recognition |
title_sort | mid level deep food part mining for food image recognition |
topic | midlevel deep FP mining food image recognition midlevel image part-based approaches deformable food parts modelling FP representations deep learning approaches |
url | https://doi.org/10.1049/iet-cvi.2016.0335 |
work_keys_str_mv | AT jiannanzheng midleveldeepfoodpartminingforfoodimagerecognition AT liangzou midleveldeepfoodpartminingforfoodimagerecognition AT zjanewang midleveldeepfoodpartminingforfoodimagerecognition |