Blockchain with Sand Cat Swarm Optimized Semantic Attentive Mobilenetv3 Model for Smart IoHT System

Nowadays, the healthcare (HC) system based on blockchain (BC) assisted internet of things (IoT) has grown a lot in which the medical services are readily accessible to each user via internet sources. However, the HC IoT systems are highly vulnerable to security threats to render the patient's personal health data. Hence, this study put forth innovative optimized Deep learning (DL) model for BC-assisted secure IoT enabled HC system. The developed study undergoes three major stages namely, medical image encryption, optimal key selection, decryption, and disease classification. Initially, the medical image is encrypted and followed by this, optimal key selection process is performed using a novel sand cat swarm optimizer (SCSO). The encrypted medical data is then stored in the BC and decrypted for further process. Finally, semantic attentive mobileNetV3 (SAtt-MNetV3) model is introduced to extract the features and to classify the decrypted health data to the patients. The simulation is carried out on the Python platform and a publicly available ISIC-2017 database is considered for training process. Various existing studies are also compared with the developed model by computing several performance measures like accuracy, computation time, peak to signal noise ratio (PSNR), recall, sensitivity and specificity metrics. The developed methodology obtained the overall accuracy of
$99.2{{\% }}$
, recall of
$98.9\%$
, specificity of
$99.5\%$
, PSNR of
$60.7{{\ dB}}$
and computation time of
$0.13{{\ s}}$
. The simulation performances outperform well compared to conventional studies.