ABSTRACT Magnetic Resonance Imaging (MRI) is a widely used non‐invasive medical imaging technique for detecting brain abnormalities, including tumors. The underlying physics of MRI originates from principles of quantum mechanics, which govern the behavior of hydrogen nuclei during image acquisition. Inspired by these quantum principles, particularly the notions of uncertainty and probabilistic representation, this work adopts a quantum‐inspired conceptual framework to address the ambiguity and noise commonly present in MRI images. In practice, MRI images often suffer from haziness and reduced clarity due to imaging noise, environmental disturbances, and limitations of acquisition devices, which complicate accurate tumor detection and segmentation. To mitigate these challenges, this study proposes a fuzzy‐based image enhancement and segmentation approach guided by quantum‐inspired concepts of uncertainty representation. The method begins by cropping the region of interest and fuzzifying the image to highlight the tumor area. The fuzzified image is then transformed into an Interval Type‐2 fuzzy set using the Hamacher T‐conorm to better model uncertainty in pixel membership values. Noise effects are further addressed by incorporating neighborhood pixel information to refine the representation. Subsequently, the processed image is segmented using the Fuzzy C‐Means clustering algorithm with a Manhattan distance metric to accurately delineate tumor regions. The proposed algorithm is evaluated using thirty (30) brain tumor MRI images and compared with nine existing clustering‐based segmentation methods through both qualitative and quantitative analyses. To evaluate the effectiveness of algorithms in adverse noise conditions, the images are additionally corrupted by introducing 6% Salt‐and‐Pepper and 2% Gaussian noise into the MRI images. In addition to that Statistical analysis is also performed using Friedman and Nemenyi test. Experimental results and statistical analysis demonstrate that the proposed approach achieves superior segmentation performance and improved noise resilience compared with the benchmark techniques, producing more accurate tumor detection even under highly noisy imaging conditions.
Rathee et al. (Wed,) studied this question.