In this paper, mutlimodal approach has been used for segmenting Tumor core in MRI images using T1-Contrast enhanced, T2-weighted, and FLAIR imaging modalities. Segmentation techniques working on single modality fails to segment brain tumor where the contrast in MR images is low or when sufficient disparity is not present in the intensity of tumor and background region. The proposed method overcome above stated problems by fusing the images of three modalities to form one image. The entire process is divided into five stages: Image Acquisition, Preprocessing, Segmentation, Tumor Extraction, and Evaluation. Two separate segmentation algorithms, Fuzzy C Means and K Means have been used. The results were evaluated using manually segmented Ground Truth. The average Dice accuracy for 18 real tumors (including 12 high grade Glioma and 06 low grade Glioma) is 86 % using Fuzzy C Means as well as K Means. Hence, the proposed method is highly efficient in segmenting tumor core.
CITATION STYLE
Goel, S., Sehgal, A., Mangipudi, P., & Mehra, A. (2017). Brain tumor segmentation in Glioma images using multimodal MR imagery. In Advances in Intelligent Systems and Computing (Vol. 479, pp. 733–739). Springer Verlag. https://doi.org/10.1007/978-981-10-1708-7_84
Mendeley helps you to discover research relevant for your work.