Enhancing indoor monitoring of visually impaired people using temporal convolutional network with optimization model in IoT environment

在物联网环境下,利用时间卷积网络和优化模型增强对视障人士的室内监测

阅读:1

Abstract

The Internet of Things (IoT) has emerged as a powerful technology in various fields, including healthcare, assisting the elderly and disabled individuals. Solution-based IoT is widely utilized in healthcare support in diverse aspects of their daily lives and activities. These IoT-based assistive systems are crucial in uplifting the quality of life for disabled individuals and older adults. Furthermore, IoT-based interior navigation methods assist with easy accessibility and movement in areas, ensuring that disabled people can find objects with greater ease. Machine learning (ML) and deep learning (DL) approaches are efficiently beneficial for detecting indoor objects for visually challenged people. This paper proposes the Metaheuristic Optimization-Driven Attention Mechanism for Enhancing Indoor Monitoring of Visually Impaired People (MOAM-EIMVIP) technique. The main intention of the MOAM-EIMVIP technique is to develop an IoT-based DL model for indoor monitoring of disabled individuals. Initially, the data pre-processing stage applies min-max normalization to convert input data into a suitable format. The sine-cosine algorithm (SCA) is employed to select the feature process to detect and choose the most relevant features from input data. Moreover, the temporal convolutional network-attention mechanism (TCNA) model is implemented for classification. Finally, the Eurasian oystercatcher optimizer (EOO) method alters the hyperparameter tuning of the TCNA model optimally and results in higher classification execution. An extensive simulation is performed to highlight the performance of the MOAM-EIMVIP approach under the HAR dataset. The experimental analysis of the MOAM-EIMVIP approach portrayed a superior accuracy value of 99.30% over existing methods.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。