Internet of agents system for age and gender classification using grasshopper optimization with deep convolution neural networks. Issue 4 (2nd August 2022)

• Record Type: Journal Article
• Authors: Dutta, Ashit Kumar; Qureshi, Basit; Albagory, Yasser; Alsanea, Majed; Gupta, Deepak; Khanna, Ashish
• Is Part Of: Expert systems. Volume 40:Issue 4(2023)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-08-02
• Subjects: age classification -- deep learning -- internet of agents -- metaheuristics -- multi agent system -- regression; Expert systems (Computer science); 006.33
• DOI: 10.1111/exsy.13115
• Languages: English
• ISSNs: 0266-4720

Abstract

Abstract: The internet of agents (IoA) is an emergent model, which mainly intends to resolve the limitations of the internet of things (IoT) devices with respect to reasoning and social competencies for improving proactivity, intelligence, and interoperability. This article presents a novel grasshopper optimization with deep learning enabled multi‐agent system for age and gender classification (GOADL‐MASAGC) model for IoA. The proposed GOADL‐MASAGC technique intends to categorize age as well as gender. The proposed GOADL‐MASAGC technique applies a multi‐agent system, which incorporates distinct processes for age and gender classification like pre‐processing, feature extraction, and classification. Besides, the GOADL‐MASAGC technique enables to performance concurrent process of classification and regression to identify age and gender respectively. In addition, the GOA with Capsule Network (CapsNet) model was executed for deriving a suitable group of feature vectors and the GOA is employed as a hyperparameter optimizer. Finally, wavelet kernel extreme learning machine (WKELM) was employed as a classifier for gender classification and deep belief network (DBN) is used as a regression approach for age recognition. For demonstrating the improved performance of the GOADL‐MASAGC model, a series of simulations were executed and the outcomes are examined in various aspects. The extensive comparative analysis reported the enhanced outcomes of the GOADL‐MASAGC approach over the existing methods.