Medical diagnostic support system based on breast thermography using Raspberry Pi and cloud computing

Abstract

Breast thermography is a promising medical imaging technique for the detection of breast cancer. However, providing a robust and portable computer-aided diagnostic system for breast thermography remains a tedious task. In this paper, a computer-aided diagnostic system based on breast thermography is developed and implemented on a Raspberry Pi 4 using the cloud computing services to provide the computing power needed for machine learning algorithms. Image processing techniques such as pre-processing and segmentation are employed to achieve an adequate feature extraction task. The Support Vector Machine classifier is used in the final stage to classify the breast as normal or abnormal. According to the experimental results, the proposed computer-aided diagnostic system has shown high performance in both the segmentation and classification steps. Furthermore, a low computation time was obtained when using the high computing capabilities of the cloud with the Raspberry Pi. We conclude that the implementation of such a decision support system on the Raspberry Pi especially when using the cloud computing services, can be a reliable tool for radiologists to predict breast abnormalities even in the rural backcountry where there is lack of health services.

Big data classification based on improved parallel k-nearest neighbor

“…When dealing with such large amounts of data, the usual performance of computers is insufficient, and more intelligence is required to keep up with the current expansion of big data [3]- [17]. High-performance computing (HPC) should be taken into account while designing and creating software since it has the potential to increase computation speed while also giving more accurate results at a lower total cost than traditional computing [18]- [25]. HPC may be classified into a variety of categories, the most prominent of which are: computer clusters, grid computing, cloud computing, graphical processing units (GPU), microprocessors, and field programmable gate array (FPGA) [4], [20]- [22], [24].…”