GourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention

We propose GourmetNet, a single-pass, end-to-end trainable network for food segmentation that achieves state-of-the-art performance. Food segmentation is an important problem as the first step for nutrition monitoring, food volume and calorie estimation. Our novel architecture incorporates both chan...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Udit Sharma, Bruno Artacho, Andreas Savakis
Formato: article
Lenguaje:EN
Publicado: MDPI AG 2021
Materias:
Acceso en línea:https://doaj.org/article/1115306637554b4f9acc4faeacaeaada
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:1115306637554b4f9acc4faeacaeaada
record_format dspace
spelling oai:doaj.org-article:1115306637554b4f9acc4faeacaeaada2021-11-25T18:56:57ZGourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention10.3390/s212275041424-8220https://doaj.org/article/1115306637554b4f9acc4faeacaeaada2021-11-01T00:00:00Zhttps://www.mdpi.com/1424-8220/21/22/7504https://doaj.org/toc/1424-8220We propose GourmetNet, a single-pass, end-to-end trainable network for food segmentation that achieves state-of-the-art performance. Food segmentation is an important problem as the first step for nutrition monitoring, food volume and calorie estimation. Our novel architecture incorporates both channel attention and spatial attention information in an expanded multi-scale feature representation using our advanced Waterfall Atrous Spatial Pooling module. GourmetNet refines the feature extraction process by merging features from multiple levels of the backbone through the two attention modules. The refined features are processed with the advanced multi-scale waterfall module that combines the benefits of cascade filtering and pyramid representations without requiring a separate decoder or post-processing. Our experiments on two food datasets show that GourmetNet significantly outperforms existing current state-of-the-art methods.Udit SharmaBruno ArtachoAndreas SavakisMDPI AGarticlesemantic segmentationfood segmentationmulti-scale featuresspatial attentionchannel attentionChemical technologyTP1-1185ENSensors, Vol 21, Iss 7504, p 7504 (2021)
institution DOAJ
collection DOAJ
language EN
topic semantic segmentation
food segmentation
multi-scale features
spatial attention
channel attention
Chemical technology
TP1-1185
spellingShingle semantic segmentation
food segmentation
multi-scale features
spatial attention
channel attention
Chemical technology
TP1-1185
Udit Sharma
Bruno Artacho
Andreas Savakis
GourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention
description We propose GourmetNet, a single-pass, end-to-end trainable network for food segmentation that achieves state-of-the-art performance. Food segmentation is an important problem as the first step for nutrition monitoring, food volume and calorie estimation. Our novel architecture incorporates both channel attention and spatial attention information in an expanded multi-scale feature representation using our advanced Waterfall Atrous Spatial Pooling module. GourmetNet refines the feature extraction process by merging features from multiple levels of the backbone through the two attention modules. The refined features are processed with the advanced multi-scale waterfall module that combines the benefits of cascade filtering and pyramid representations without requiring a separate decoder or post-processing. Our experiments on two food datasets show that GourmetNet significantly outperforms existing current state-of-the-art methods.
format article
author Udit Sharma
Bruno Artacho
Andreas Savakis
author_facet Udit Sharma
Bruno Artacho
Andreas Savakis
author_sort Udit Sharma
title GourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention
title_short GourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention
title_full GourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention
title_fullStr GourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention
title_full_unstemmed GourmetNet: Food Segmentation Using Multi-Scale Waterfall Features with Spatial and Channel Attention
title_sort gourmetnet: food segmentation using multi-scale waterfall features with spatial and channel attention
publisher MDPI AG
publishDate 2021
url https://doaj.org/article/1115306637554b4f9acc4faeacaeaada
work_keys_str_mv AT uditsharma gourmetnetfoodsegmentationusingmultiscalewaterfallfeatureswithspatialandchannelattention
AT brunoartacho gourmetnetfoodsegmentationusingmultiscalewaterfallfeatureswithspatialandchannelattention
AT andreassavakis gourmetnetfoodsegmentationusingmultiscalewaterfallfeatureswithspatialandchannelattention
_version_ 1718410506945953792