Diagnosis of liver cancer on computed tomography using medical image segmentation based on FFCM and SVM

Liver cancer is one of the most common internal malignant diseases worldwide. Computed tomography is one of the most common imaging modalities for the diagnosis of liver diseases such as hepatocellular carcinoma. Because of the high level of noise in the CT data, the contrast between the tumor and the main part of the liver shows a lower value, therefore finding and drawing a tumor will be difficult. In addition to the limitations of the imaging procedure, liver tumor classification will be complicated because of the variety of the tumor in size and structure, and that the tumors can be present virtually anywhere in the liver. This paper attempts to detect the tumor segment in images of patients with liver tumors. The proposed method for tumor classification is performed in three stages. First, pre-processing will take place on the images. In preprocessing, filters are applied to improve image quality. Then, liver boundaries are indicated in the CT image with segmentation. Segmentation is divided into two stages: border detection border extraction. the primary search area are obtained using the histogram of images and in this stage morphological operations has done using liver anatomy information about the shape and location of it, which will eventually be searched for the exact border of the liver using FFCM; in third level for cancer detection, the SVM classifier in the obtained region is used. The results of this algorithm show a precision of 90.76.

Keywords:

Computed tomography , support vector machine , Tumor segmentation , Histogram , Morphology filters.