A Lightweight Hybrid Model with Location-Preserving ViT for Efficient Food Recognition

doi:10.3390/nu16020200

CSpace

	A Lightweight Hybrid Model with Location-Preserving ViT for Efficient Food Recognition
	Sheng, Guorui 1; Min, Weiqing 2,3; Zhu, Xiangyi 1; Xu, Liang 1; Sun, Qingshuo 1; Yang, Yancun 1; Wang, Lili 1; Jiang, Shuqiang 2,3
	2024
发表期刊	NUTRIENTS
卷号	16 期号:2 页码:16
摘要	Food-image recognition plays a pivotal role in intelligent nutrition management, and lightweight recognition methods based on deep learning are crucial for enabling mobile deployment. This capability empowers individuals to effectively manage their daily diet and nutrition using devices such as smartphones. In this study, we propose an Efficient Hybrid Food Recognition Net (EHFR-Net), a novel neural network that integrates Convolutional Neural Networks (CNN) and Vision Transformer (ViT). We find that in the context of food-image recognition tasks, while ViT demonstrates superiority in extracting global information, its approach of disregarding the initial spatial information hampers its efficacy. Therefore, we designed a ViT method termed Location-Preserving Vision Transformer (LP-ViT), which retains positional information during the global information extraction process. To ensure the lightweight nature of the model, we employ an inverted residual block on the CNN side to extract local features. Global and local features are seamlessly integrated by directly summing and concatenating the outputs from the convolutional and ViT structures, resulting in the creation of a unified Hybrid Block (HBlock) in a coherent manner. Moreover, we optimize the hierarchical layout of EHFR-Net to accommodate the unique characteristics of HBlock, effectively reducing the model size. Our extensive experiments on three well-known food image-recognition datasets demonstrate the superiority of our approach. For instance, on the ETHZ Food-101 dataset, our method achieves an outstanding recognition accuracy of 90.7%, which is 3.5% higher than the state-of-the-art ViT-based lightweight network MobileViTv2 (87.2%), which has an equivalent number of parameters and calculations.
关键词	food recognition lightweight global feature ViT nutrition management
DOI	10.3390/nu16020200
收录类别	SCI
语种	英语
WOS研究方向	Nutrition & Dietetics
WOS类目	Nutrition & Dietetics
WOS记录号	WOS:001151224800001
出版者	MDPI
引用统计	被引频次：1[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/38396
专题	中国科学院计算技术研究所
通讯作者	Yang, Yancun
作者单位	1.Ludong Univ, Sch Informat & Elect Engn, Yantai 264025, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China 3.Univ Chinese Acad Sci, Sch Comp Sci & Technol, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Sheng, Guorui,Min, Weiqing,Zhu, Xiangyi,et al. A Lightweight Hybrid Model with Location-Preserving ViT for Efficient Food Recognition[J]. NUTRIENTS,2024,16(2):16.
APA	Sheng, Guorui.,Min, Weiqing.,Zhu, Xiangyi.,Xu, Liang.,Sun, Qingshuo.,...&Jiang, Shuqiang.(2024).A Lightweight Hybrid Model with Location-Preserving ViT for Efficient Food Recognition.NUTRIENTS,16(2),16.
MLA	Sheng, Guorui,et al."A Lightweight Hybrid Model with Location-Preserving ViT for Efficient Food Recognition".NUTRIENTS 16.2(2024):16.