Internet of things enabled framework for terahertz and infrared cancer imaging

The fourth industrial revolution (4IR or Industry 4.0), its communication and information enabling technologies are revolutionizing the production and services applications. In the healthcare domain, electronic health (eHealth) is evolving towards healthcare 4.0 through the Internet of things, big data analytics, Cloud and Fog computing technologies. There is need for early cancer detection, which can be achieved by new imaging technologies for example Terahertz imaging and Infrared thermography. The integration of such new imaging modalities and wearable sensor networks with computer aided diagnosis systems (CAD) and IoT systems with interconnectivity and interoperability could enable continuous, remote observation of cancer diagnosed patients by experts to monitor the success of treatment procedures, side effects of drugs, statistical analysis and making predictions of cancer based on age and demographics etc. for timely intervention and personalized care. In the healthcare 4.0 paradigm, there is need not only for smart diagnosis but also care through remote patient monitoring. The vital signs associated with cancer and clinical information can be integrated to make the system culminate into a smart system with health records. This work proposes an IoT enabled deep learning framework for classification of breast images; we propose the development of both custom CNN model and fine-tuned, pretrained CNN models capable of multimodal image classification. IoT simulation and cloud processing is performed in the ThingSpeak cloud platform with email alert capability. The performance of the CNN models has been evaluated using the accuracy, sensitivity and specificity metrics which achieved 98.44%, 98.44% and 98.45% respectively for the custom CNN model together with the receiver operating characteristics curves and confusion matrices.