Neutrosophic-Based Feature Set (NBFS) For Brain Tumor Detection Using  GLCM Features and KNN Classifier On Mri Images

S. Gowri; V. M Vijayalakshmi

Authors

S. Gowri Avinashilingam Institute for Home Science and Higher Education for Women. Coimbatore. Tamil Nadu. India;
V. M Vijayalakshmi Avinashilingam Institute for Home Science and Higher Education for Women. Coimbatore. Tamil Nadu. India;

Keywords:

Brain MRI, Brain Tumor Detection, Neutrosophic Set, Neutrosophic-Based Feature Set (NBFS), Truth, Indeterminacy, Falsehood (T, I, F), Texture Feature Extraction, Gray Level Cooccurrence Matrix (GLCM), K-Nearest Neighbor (KNN) Classifier, Medical Image Processing, Computer-Aided Diagnosis (CAD), Noise Robustness, Uncertainty Handling.

Abstract

Brain tumor detection using Magnetic Resonance Imaging (MRI) is essential for early
diagnosis and treatment planning. However, MRI images often contain noise, uncertainty, and
indistinct tumor boundaries, which challenge the effectiveness of traditional feature extraction and
classification techniques. Many existing methods fail to handle ambiguous regions robustly and
overlapping tissue characteristics, leading to decreased diagnostic reliability. To address these
limitations, this study proposes a novel computer-aided diagnosis framework based on Neutrosophic
Bipolar Fuzzy Set (NBFS) theory, which integrates neutrosophic logic with bipolar fuzzy reasoning.
This approach enables the simultaneous representation of both positive and negative degrees of
truth, indeterminacy, and falsity, improving the modeling of complex and uncertain regions in brain
MRI scans. The images are transformed into the NBFS domain, and texture features are extracted
from the Truth (T), Indeterminacy (I), and Falsity (F) components using the Gray Level Co-occurrence
Matrix (GLCM). These features are used both individually and in combination to train a K-Nearest
Neighbor (KNN) classifier. Experimental results demonstrate that the NBFS-based framework
achieves higher classification accuracy, sensitivity, specificity, and precision compared to
conventional texture-based approaches. The confusion matrix analysis further confirms reduced
misclassification rates, highlighting the robustness of the method. These findings establish the NBFS
framework as a promising tool for improving brain tumor detection in clinical decision support
systems, especially under uncertain imaging conditions.