Recent Papers

Purpose the intrinsically heterogeneous signal characteristics of a brain tumor like glioblastoma make it challenging to detect and segment the tumor in MR images. Brain tumor MRI scans were segmented using a robust method that was developed and tested. Techniques Basic edges and factual strategies can't sufficiently portion the different components of the GBM, like neighbourhood contrast upgrade, rot, and edema. The majority of voxel-based methods fail to deliver satisfactory results when applied to larger data sets, and generative and discriminative model-based methods have inherent application limitations, such as limited sample set learning and transfer. By collecting and analyzing a large amount of data, these two projects promised to model the complex interaction between the brain and behavior as well as comprehend and diagnose brain diseases. Major obstacles arose when it came to sharing, analyzing, and archiving the expanding datasets from neuroimaging. In the field of Big Data, new technologies and computational methods have emerged, but they have not yet been fully adapted for use in neuroimaging. In this work, we present the ongoing difficulties of neuroimaging in a major information setting. We review our efforts to develop a data management system for the large-scale fMRI datasets and present our novel algorithms and methods. A novel approach was developed to address these issues. Algorithms are used to break up multimodal MR images into super pixels in order to solve the problem of sampling and make the sample more representative. Multilevel Gabor wavelet filters were then used to extract features from the super pixels. To overcome the limitations of previous generative models, a conditional Random Field Grey Level Co-occurrence Matrix (GLCM) model and an affinity metric model for tumours were trained on the features. Conditional random fields theory was used to segment the tumor in a maximum a posteriori fashion using the smoothness prior defined by our affinity model and the output of the Grey Level Co-Occurrence Matrix (GLCM) and spatial affinity models. At long last, naming clamor was eliminated utilizing "primary information" like the balanced and nonstop qualities of the cancer in spatial area. Finally, "structural knowledge" like the tumours symmetrical and continuous spatial characteristics were used to eliminate labelling noise. On augmented images, the (Bat Algorithm) models were trained, tested, and validated.

Copyright © 2023 Saveetha P, Dharanya A, Pavithra K S, Vaishnavi S. This is an open access article distributed under the Creative Commons Attribution License.