A MODIFIED FUZZY C MEAN-BASED LIVER TUMOR SEGMENTATION

Abstract

Early detection of the liver cancer disease prediction reduces the risk factor of the patients. Several conventional segmentation techniques are proposed to detect the liver tumor like Gaussian mixture model, NS, FCM, and Watershed. But most of the Conventional models difficult to detect liver tumor in the early stages and to improve the life of patients segmenting the liver tumor is essential in detecting and managing liver cancer. This study suggests using the modified fuzzy c means (MFCM) algorithm to segment liver Tumors. In this work, proposed a modified Fuzzy C Mean algorithm to do the liver segments depends on the region of interest over the filtered image. The suggested technique enhances the segmentation accuracy, and robustness by including spatial information in the conventional fuzzy c means algorithm. The performance of the suggested algorithm was assessed on a dataset of CT scans of the liver that was made available to the public. The results demonstrate that the MFCM algorithm surpasses the conventional fuzzy c means technique and several other advanced segmentation approaches regarding segmentation accuracy, sensitivity, specificity, and dice similarity coefficient. Clinicians can benefit significantly from the proposed algorithm's assistance in detecting and managing liver cancer. The results show that the suggested model works, which outperformed many existing algorithms, attained 98% accuracy.