An Innovative Neutrosophic Combinatorial Approach Towards the Fusion and Edge Detection of MR Brain Medical Images

Abstract

This research proposes the idea of implementing an innovative mechanism to detect the edges in distinct MR brain medical images based on the aspect of Neutrosophic sets (NSs). NS-based entropy is one of the emerging tools to procure a neutrosophic image from the crisp image. Followed by the aforementioned procedure, fusion has been done for the neutrosophic image in order to acquire fused neutrosophic images (FNI) then the FNI’s are again regenerated to form a fused crisp image. Later, the Bell-Shaped (BS) function and the Sobel operator works on the FNI to obtain a combination of three subsets. After determining the neutrosophic subsets, various entropies such as Norm, Threshold, Sure, and Shannon act on it to provide their threshold values, and the computed subsets along with thresholds are incorporated to produce a new binarized image. Subsequently, morphological operations were implemented to construct the image edges. The resultant images with different entropies are compared by using the performance measurement factors. Based on the measurement factors, the proposed Norm entropy image edge detection innovations have proven to be an efficient tool with reference to other entropies. In addition, the Norm entropy-based proposed method was compared with some of the other existing edge detection methods inclusive of Sobel, Chan, Tian, and Wu. FOM and PSNR factors have been applied to estimate the results of edge detection achieved through five distinct methods. The findings confirmed that the implementation of the proposed object edge detection mechanism is much stronger compared to other existing methods.