Due to the modern technology and the importance of image processing and digital signals, using image processing and segmentation algorithms are very useful to be used in medical issues and diseases. For healthcare, Medical Segmentation Decathlon (MSD) made a challenge for different tasks to test the mechanisms and algorithms of image processing and to test the machine. In this dataset, computed tomography (CT) scan had been used to segment the liver tumours. The challenge in this dataset is to distinguish between large liver and small tumours. The size is 201 3D volumes (131 Training + 70 Testing). The source is IRCAD H?pitaux Universitaires. They did a 3D scan for 20 cases half of them 10 men and half are women. Most of them have tumours in their livers. They are around 75% of them.
The challenge studies the size of liver (width, depth, height). It also studies the location of tumours for each patient. This indicate the impacts of other organs on liver shape and density.
25% of the cases have no tumours. Dataset shows that men have more liver tumours than female. In addition, the patients who are older they are able to get more liver tumours comparing to the younger ones.
Liver cancer is one of the most common types of cancer. Cancer is a type of tumours, which they are amount of cells growing and behaving in abnormal way. They could be cancerous or they could be un cancerous type. Liver segmentation is one of the hardest segmentations because of several variables because of its shape and many organs closer. Also, it contains some other diverse pathologies like fat and etc. Liver’s normal shape and size is different from one patient to another patient. Specialist can calculate if it is normal or abnormal, but it becomes easier with CT to evaluate the size, shape and cells. Based on this fact and issue, specialists found a way to segment the liver to look whether all cells of the liver are normal or there are some group of cells are tumours. Computed tomography (CT) is an image process used for highly detection of tumours. CT scan (image processing) take the important information of liver and some algorithms to calculate the signals of the images one by one. It starts scanning the liver from one side across all parts of liver to get the full image. Then, do some operations and process of each part to detect the tumours. The basic idea behind the CT is the value of the gray color of liver. Depending on grayscale the rate of gray color is between (0 and 225). The normal color of the liver should have a value between (90 and 92). If the value of gray is exceeding this amount it will be classified as abnormal cells (tumours) . Also, it records the value of each pixel, shape and the smooth of liver and tests the data by comparing it with the normal standard information of normal liver. But gray scale image is not accurate because of soft tissue in liver that cause overlapping in CT. Specialists can segment the small and big tumor using imaging and algorithms in order to detect it earlier and avoid reaching the level of death, so the challenge here is early detection of tumours specially the cancer.
In order to detect the tumours and segment them earlier to give the patients the best medical treatment, specialists develop the CT process. It is important and challenge to understand the structure of liver, the blood vessels, shape and the relationship between them. Algorithms of different methods have been developed to help computed tomography (CT) to build a 3D images of liver. In 3D image of liver, they can create a stander shape and size of liver, and compare the image that may has tumours to detect them by point-to-point method. Most important two approaches are contour-based and pixel-based. First important step for segmentation is to find the problem with the tissue part of liver where there is a difference between this part and other surface of it.
The important of liver segmentation whether for normal or abnormal liver are surgery for donor liver, treatments and study the size and structure of liver. First, specialists develop algorithm using 3D to detect liver and segmentation. Second, they develop another algorithm to divide the image to many parts.