Artificial Intelligence with Optimization Algorithm Based Robust Sign Gesture Recognition for Visually Impaired People
Ameer N. Onaizah · International Journal on Computational Modelling Applications · 2024
Sign language is an effective means of communication with visually impaired individuals, as it can be utilized anywhere. So, gestures play a significant part in communication between deaf and dumb people. They are a form of non-verbal data exchange that has garnered significant attention in the development of Human-Computer Interaction (HCI) models, as they permit consumers to statethemselves intuitively and naturally in dissimilar contexts. Sign gesture detection is the main need of any HCI application, e.g. gaming, virtual reality, and monitoring methods. At present, Computer vision (CV) and artificial intelligence (AI) have been the efficient areas of dynamic research and growth with the advances in the assistive technology field. In this manuscript, we design and develop an Enhanced Sign Gesture Recognition Model for Disabled People Using Advanced Optimization Models (ESGRM-DPAOM). The proposed ESGRM-DPAOM system is to improve the sign gesture recognition solutions for visually impaired individuals. To accomplish that, the proposed ESGRM-DPAOM model initially applies image preprocessing using wiener filtering (WF) to eliminate the noise in input image data. For the feature extraction process, the SqueezeNet model has been employed and an optimal parameter tuning model utilizes pigeon-inspired optimization (PIO). In addition, the proposed ESGRM-DPAOM models involve the classification processwith the aid of the elman neural network (ENN) model. At last, the golden jackal optimizer (GJO) algorithm adjusts the hyperparameter values of the ENN model optimally and outcomes in greater classification performance. Extensive experimentation led to authorizing the performance of the ESGRM-DPAOM approach. The simulation outcomes specified that the ESGRM-DPAOM system emphasizedadvancement over other existing methods.