Smart IoMT-Based Disease Prediction Framework Using Deep Learning Techniques with Fractional-Order Hyperchaotic Encryption Model

Information technology has enhanced the use of the Internet of Things (IoT). In general, IoT impacts various sectors including data management, smart cities, and the medical profession. The sustainable medical system has developed significantly using the Internet of Medical Things (IoMT). In the healthcare sector, it has gained a huge impact in validating and monitoring the information about the patient. Because of the sudden expansion of IoMT, it is also essential to secure sensitive medical information. For personal health surveillance, the investigation developed an IoMT-based monitoring. Due to these limitations, a novel approach is developed in this research work. At first, the needed medical images are collected from the benchmark resources. The collected image is encrypted using a hybrid encryption (HyE), which is an integration of the fractional-order hyperchaotic system (FOHS) and advanced encryption standard (AES). The FOHS makes encryption along with the AES for enhancing security because it makes the pseudorandom sequence more complex. The encrypted images are decrypted for disease prediction to minimize the mortality rate. In this proposed model, an Adaptive Dilated InceptionNet with Sparse Attention (ADINet-SA) is developed for the final forecast task. The dependability of this proposed deep learning-based disease prediction is enhanced by tuning the parameters in ADINet-SA using the Boolean derived sculptor optimization (BDSO). In the end, the outstanding functionality is validated and compared with previous baseline frameworks. The overall disease detection process of the implemented method has achieved 97% accuracy, 98% precision, 94% specificity, 92% sensitivity, and 94% NPV in dataset 1. In dataset 2, the overall performance of the suggested model attains 93% accuracy, 92.4% sensitivity, 94% specificity, 93% precision, and 94% NPV. This performance enhancement helps to improve the disease detection process effectively.