Answers are grounded in this paper and cite their sources.
Observational analysis improved sign language recognition accuracy by 93.07% in users with hearing disabilities, indicating deep learning's potential in communication enhancement.
Sign language recognition (SLR) was employed in human-machine interactions, enhancing the survival of people with speech losses or who rely on non-verbal cues. The Internet of Things (IoT), the world of sensors, might reduce the absence of the hearing sense class of the physically challenged. IoT's explosive expansion can help those with hearing impairments by enhancing their everyday activities, communication, access to education, and general quality of life, both for them and their caretakers. The usage of DL techniques can aid in decreasing barriers to communication. This manuscript proposes an Enhancing Communications with Others through Robust Sign Language Recognition using the Sparse Autoencoder with Hearing Disabilities (ECRSLR-SAEHD) approach. The presented ECRSLR-SAEHD approach utilises a hyperparameter-tuned deep learning model to detect SL and leverages the IoT to assist people with hearing disabilities. Initially, the Gaussian filtering (GF) method is utilised for the image preprocessing process to enhance image clarity by reducing noise. For the feature extraction process, the EfficientNetB7 method is used to capture complex image patterns for a robust representation of complex SL movements. For the recognition of SL, the sparse auto-encoder (SAE) model is employed. To improve the SL performance of the SAE model, the fennec fox algorithm (FFA) model is used for the hyperparameter tuning process. To demonstrate the good classification outcome of the ECRSLR-SAEHD technique, a wide range of simulations is accomplished under a benchmark dataset. A widespread experimental validation of the ECRSLR-SAEHD technique portrayed a superior accuracy value of 93.07% over existing methods.
No takes yet. Share an insight, caveat, or question.
Maashi et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: