Machine Learning Approach-Based Gamma Distribution for Brain Tumor Detection and Data Sample Imbalance Analysis

Abstract

Recently, artificial intelligence applications in magnetic resonance imaging have been applied in several clinical studies. The analysis of brain tumors without human intervention is considered a significant area of research because the extracted brain images need to be optimized using a segmentation algorithm that is highly resilient to noise and cluster size sensitivity problems and automatically detects the region of interest (ROI). In this paper, an improved orthogonal gamma distribution-based machine-learning approach is used to analyze the under-segments and over-segments of brain tumor regions to automatically detect abnormalities in the ROI. Further data imbalances due to improper edge matching in the abnormal region is sampled by matching the edge coordinates and sensitivity, and the selectivity parameters are measured using the machine learning algorithm. The benchmark medical image database was collected and analyzed to validate the efficiency and accuracy of the optimal automatic detection in tumor and non-tumor regions. The mean error rate of the algorithm was determined using a mathematical formulation. The system is evaluated based on experimental results that showed the method of orthogonal gamma distribution with the machine learning approach attained an accuracy of 99.55% in detecting brain tumors. This research contributes to the field of brain abnormality detection and analysis without human intervention in the health care sector. INDEX TERMS Magnetic resonance imaging, gamma distribution, machine-learning algorithm, brain abnormality.

Brain–computer interface for assessment of mental efforts in e‐learning using the nonmarkovian queueing model

“…The comparison of classification accuracies in terms of confusion matrix with 10‐fold cross‐validation for RF, SVM, and KNN is presented in Tables 2–4. In the tables, A represents the output of “can understand” class and B represents the output of “cannot understand” cases [19].…”

Deep Learning-Based HCNN and CRF-RRNN Model for Brain Tumor Segmentation

“…Therefore, only a few characteristics were obtained [26] and very low precision was achieved in tumor detection. Furthermore, the ELM-LRF,ConvNets,DLF, MLPMis lacking with the overlap measure [27], [26], which is a dix-sequence index, an important parameter for assessing the exactness of any brain tumor segmentation algorithm.…”

L1 norm based pedestrian detection using video analytics technique

“…Here the researcher 13 uses the ensemble approach by combining background subtraction and HOG for feature extraction. The approach has the ability to get accurate results for an image but fails in video due to a dynamic modification in video illumination and background, hence not updating the background.…”