The detection of a brain tumor and stroke from the magnetic resonance imaging (MRI) is one of the... more The detection of a brain tumor and stroke from the magnetic resonance imaging (MRI) is one of the critical tasks in recent days for neuro-radiologists. So, various segmentation techniques are developed customarily, but it fails to provide an accurate diagnosis. To elucidate this problem, this paper aims to develop a fusion based segmentation technique for the detection of MRI brain tumor and stroke. The MRI brain images considered here include T1-weighted (T1-w), T2-weighted (T2-w), Diffusion Weighted Imaging (DWI), and Fluid-attenuated Inversion Recovery (FLAIR). The first step in the proposed methodology includes Gradient based Discrete Wavelet Transform as an image fusion technique with the target gradient estimation process. The different image fusion combinations include T1-w and T2-w, T1-w and DWI, T1-w and FLAIR, T2-w and DWI, T2-w and FLAIR, DWI and FLAIR. Secondly, the visual quality of the image is improved by applying the histogram equalization method. Finally, an Intensity Factorized Thresholding technique is proposed for segmentation in order to emphasis the diagnosis of tumor and stroke affected region in the given MRI brain image based on the pixel intensity. Here, the segmented results of both original (non-fused) and fused images are evaluated for predicting the accurate region of tumor and stroke. During experiments, the performance of both existing and proposed techniques are evaluated by using various measures like sensitivity, specificity, accuracy, Positive Predictive Value, Negative Predictive Value, Rand Index, Global Consistency Error, Variation of Information, Jaccard and Dice coefficients. From the obtained result, it is concluded that fusion based segmentation technique is giving better results than non-fusion based segmentation techniques. Among the fusion based segmented result, T2-w and FLAIR fused segmented result is superior to other fusion combinations for detecting tumor. Similarly, DWI and FLAIR fused segmented result is better than other fusion combinations for diagnosing stroke.
The detection of a brain tumor and stroke from the magnetic resonance imaging (MRI) is one of the... more The detection of a brain tumor and stroke from the magnetic resonance imaging (MRI) is one of the critical tasks in recent days for neuro-radiologists. So, various segmentation techniques are developed customarily, but it fails to provide an accurate diagnosis. To elucidate this problem, this paper aims to develop a fusion based segmentation technique for the detection of MRI brain tumor and stroke. The MRI brain images considered here include T1-weighted (T1-w), T2-weighted (T2-w), Diffusion Weighted Imaging (DWI), and Fluid-attenuated Inversion Recovery (FLAIR). The first step in the proposed methodology includes Gradient based Discrete Wavelet Transform as an image fusion technique with the target gradient estimation process. The different image fusion combinations include T1-w and T2-w, T1-w and DWI, T1-w and FLAIR, T2-w and DWI, T2-w and FLAIR, DWI and FLAIR. Secondly, the visual quality of the image is improved by applying the histogram equalization method. Finally, an Intensity Factorized Thresholding technique is proposed for segmentation in order to emphasis the diagnosis of tumor and stroke affected region in the given MRI brain image based on the pixel intensity. Here, the segmented results of both original (non-fused) and fused images are evaluated for predicting the accurate region of tumor and stroke. During experiments, the performance of both existing and proposed techniques are evaluated by using various measures like sensitivity, specificity, accuracy, Positive Predictive Value, Negative Predictive Value, Rand Index, Global Consistency Error, Variation of Information, Jaccard and Dice coefficients. From the obtained result, it is concluded that fusion based segmentation technique is giving better results than non-fusion based segmentation techniques. Among the fusion based segmented result, T2-w and FLAIR fused segmented result is superior to other fusion combinations for detecting tumor. Similarly, DWI and FLAIR fused segmented result is better than other fusion combinations for diagnosing stroke.
Uploads
Papers by Deepa Bharathi