Algoritmo de segmentação do fígado em tomografias através de técnicas de processamento digital de imagens

Abstract

The liver plays an important role in the harmonious functioning of the human body. For liver complications that require surgical interventions, the location and volume of the organ are analyzed by dynamic computed tomography, due to the short acquisition time, wide image range and high spatial resolution. Despite the advantages, it presents numerous difficulties due to the physical characteristics of the region. For example, in the CT scan, there are several adjacent organs, such as the pancreas and the stomach, which have similar intensity levels to the liver. Thus, this study aimed to perform the segmentation of the liver with the aid of medical image processing techniques in CT scans of the upper abdomen to implement the calculation of liver volume and visualization of the structure in three dimensions, due to its importance in the planning of surgeries, mainly to analyze possible donors for liver transplantation, where knowledge of the volume and location enables more effective treatments and surgical procedures. In image processing, segmentation and unsupervised learning techniques were applied, obtaining an average success rate of 87.77% and a maximum of 98.23%.