Contents I. Introduction
Some chronic diseases are linked to a higher risk that stems from Visceral Adipose Tissue (VAT) in the abdomen. Imagining methods have been used to get images of parts of or all of the human body. Those images are used to establish how much fat is in the body and how it is spread and distributed. Dual energy X-ray absorptiometry (DEXA), air displacement plethysmography, hydrostatic weighing, hydrodensitometry, and bioelectrical impedance (BIA) are some of the methods which are exploited in body fat assessment. In the last couple of decades there has been an increase in using the Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) scans. Medical imaging methods that shows a cross section of the body can distinguish between different types of fat such as visceral and subcutaneous. One advantage of the MRI over the CT scan is that it is done without having any harmful long term effects that are associated with radiation. Therefore, in many studies, the choice of MRI to be involved is justified from a patient health and safety point of view. This is particularly regarded when clinicians are considering studies over a long period of time as well as other studies that may include infants, young kids or adolescents. The authors will suggest to use a new convolutional neural network technique to assist in SAT and VAT quantification employing images acquired and enhanced. The hope is that with the use of artificial intelligence, the results will not be dependent on who did the work and it can be done separately at different locations and produce the same results. The paper will have the following format; the background is described first. This is followed by referring to related research work on the same area. Subsequently, the collection and cleaning of the data is presented. The new method which used convolutional neural networks is then introduced. The results of applying the new technique is then shown. Section VII summarizes the work and proposes future developments.
