Training deep neural networks for wireless sensor networks using loosely and weakly labeled images

NEUROCOMPUTING

卷:427

Article

Zhou, Qianwei^1,2  Chen, Yuhang^1,2  Li, Baoqing³  Li, Xiaoxin^1,2  Zhou, Chen^1,2  Huang, Jingchang³  Hu, Haigen^1,2 

[1] Zhejiang Univ Technol, Coll Comp Sci & Technol, Hangzhou 310023, Peoples R China

[2] Key Lab Visual Media Intelligent Proc Technol Zhe, Hangzhou 310023, Peoples R China

[3] Chinese Acad Sci, Shanghai Inst Microsyst & Informat Technol, Shanghai 200050, Peoples R China

关键词: Deep neural networks;    Wireless sensor networks;    Automated data labeling;    Image recognition;    Transfer learning;    Model compression;   

DOI  :  10.1016/j.neucom.2020.09.040

来源: Elsevier

【 摘 要 】

Although deep learning has achieved remarkable successes over the past years, few reports have been published about applying deep neural networks to Wireless Sensor Networks (WSNs) for image targets recognition where data, energy, computation resources are limited. In this work, a Cost-Effective Domain Generalization (CEDG) algorithm has been proposed to train an efficient network with minimum labor requirements. CEDG transfers networks from a publicly available source domain to an application specific target domain through an automatically allocated synthetic domain. The target domain is isolated from parameters tuning and used for model selection and testing only. The target domain is significantly different from the source domain because it has new target categories and is consisted of low quality images that are out of focus, low in resolution, low in illumination, low in photographing angle. The trained network has about 7 M (ResNet-20 is about 41 M) multiplications per prediction that is small enough to allow a digital signal processor chip to do real-time recognitions in our WSN. The category level averaged error on the unseen and unbalanced target domain has been decreased by 41.12%. (c) 2020 Published by Elsevier B.V.

【 授权许可】

Free