Multiregional Segmentation Modeling in Medical Ultrasonography: Extraction, Modeling and Quantification of Skin Layers and Hypertrophic Scars

Abstract

In the clinical practice of the burns treatment, an autonomous modeling of the burns morphological structure is important for a correct diagnosis. Unfortunately, the geometrical parameters of burns and skin layers are subjectively estimated. This approach leads to the inaccurate assessment depending on the experience of an individual physician. In our research, we propose the analysis of multiregional segmentation method which is able to differentiate individual skin layers in the ultrasound image records. The segmentation method is represented by the mathematical model of skin layers while other structures are suppressed. Skin layers are consequently approximated by their skeleton with target of the layers distance measurement. The main applicable output of our research is the clinical SW SkinessMeter 1.0.0 serving for an autonomous modeling and quantification of the skin layers.