Browsing by Author "Rajagopalan, V"

Now showing 1 - 2 of 2
  • Item
    Brain Tumor Segmentation by Integrating Symmetric Property with Region Growing Approach
    (2015 ANNUAL IEEE INDIA CONFERENCE (INDICON), 2015) Gupta, M; Gayatri, KS; Harika, K; Rao, BVVSNP; Rajagopalan, V; Das, A; Kesavadas, C
    Brain tumor segmentation is an important procedure for early diagnosis of brain tumor and planning of its treatment. However it is still a difficult task due to variations in size, shape and location of tumor. In this paper, we propose a novel brain tumor segmentation method using T2-weighted brain MR images by integrating symmetry property of brain with region growing approach. Bilateral symmetry property of brain is used in our method to identify various regions having probability of presence of the tumor. Identification of exact tumor location and its segmentation is then performed by using region growing technique. Qualitative and quantitative evaluation of proposed approach was performed and promising results have been demonstrated when compared with ground truth and other state of art method. The segmented tumor region obtained in our work can assist the doctors and radiologist in the diagnosis of brain tumor and treatment planning.
  • Item
    Volumetric Segmentation of Brain Tumor Based on Intensity Features of Multimodality Magnetic Resonance Imaging
    (2015 INTERNATIONAL CONFERENCE ON COMPUTER, COMMUNICATION AND CONTROL (IC4), 2015) Gupta, M; Rao, BVVSNP; Rajagopalan, V; Das, A; Kesavadas, C
    Brain tumor segmentation is a very difficult task due to vast diversity in appearance of tumor tissues among different patients and huge variations in their size and shape. In this paper, we proposed an approach which focuses on volumetric segmentation of brain tumor in three dimensions using multiple modalities of MRI brain volume including T1-weighted, T2-weighted and T1-contrast enhanced data set. Since these modalities have different contrasts for same tissue type we combined them to develop a robust segmentation approach. Qualitative and quantitative assessment of our segmentation algorithm was performed which demonstrates that proposed algorithm performs well in comparison with ground truth and other state of art methods. Proposed work provides 3D visualization of segmented tumor volume which can assist the doctors and radiologist in the diagnosis of brain tumor, as it gives information about tumor location and its features.