Open AccessSegmentation of Tumor in MRI Brain Images using Morphological Operators and Non-Local Means Filter
Author(s) -
Kavin Kumar K,
Meera Devi T,
R. Abirami,
R. Akila,
D Ashok
Publication year - 2019
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.d4459.118419
Subject(s) - rician fading , artificial intelligence , noise (video) , filter (signal processing) , segmentation , pattern recognition (psychology) , computer science , thresholding , noise reduction , median filter , peak signal to noise ratio , magnetic resonance imaging , metric (unit) , computer vision , image processing , image (mathematics) , medicine , radiology , algorithm , operations management , decoding methods , fading , economics
Brain Tumor is the abnormal development of tissues in the brain. According to survey report Times of India, 2019 around 5, 00,000 people are diagnosed with brain tumor in India. Among 5, 00,000 people 20 percent are children. Magnetic resonance image (MRI) used for clinical analysis of human body are sensitive to redundant Rician noise. Rician is the type of noise added during the acquisition of MRI. The removal of noise variance can be performed by constructing many filters. Among those filters, non-local means filter is used for denoising the Rician noise. In this project simulated MRI data and real time clinical data of T1, T2 and Proton Density weighted MRI images are de-noised and the performance metrics is analyzed using PSNR (Peak Signal to Noise Ratio) and SSIM (Structural Similarity Index Metric). The de-noised image is then subjected to thresholding and morphological operators and the tumor region is segmented.