Human Tumor Detection Using Active Contour and Region Growing Segmentation

Segmentation and detection of borders are one of the most important task in image processing. The segmented images are divided into a different range of areas. In many image processing programs, this is an intermediate step, such as content-based image and video recovery, object detection and tracking, biomedical engineering, and many others. The segmentation of human brain tumors from data sets of magnetic resonance imaging (MRI) is of first-rate importance for advanced analysis, prediction of rate increases and remedies. However, automating this method is hard because of the existence of extreme partial quantity effect and high-quality variability in tumor structures. In this work, we introduced the area of developing segmentation and active contour model for the automatic segmentation of the human brain tumor from multimodal volumetric MRI images. Using active contour and region growing segmentation, we progressed the accuracy and robustness to detect more than one brain tumor. With the accurate place and shape approximation within the initial degree, the tumor is effectively detected with the aid of using active contour. The correctness and robustness of the proposed algorithm are verified on extracting the tumor areas of MRI. Compared with the tumor outlines, using the outlines based on the proposed set of rules can achieve sensitivity values of up to 98.98%. The proposed algorithm presents an easy and accurate approach for tumor segmentation in MRI images.