Sequence of image enhancemant of flat electroencephalography using intuitionistic fuzzy set

Abstract

This study focused on contrast enhancement of Flat Electroencephalography (fEEG) image during epileptic seizure. The main interest is in visualizing the path of brainstorm in the brain that occur during seizure. Selected techniques that are involved ranging from classical, ordinary fuzzy, and advanced fuzzy namely the intuitionistic fuzzy sets (IFS). Different techniques may result in different output of fEEG image. The methods in classical approach are Power Law Transformation, Histogram Equalization, and Image Size Dependent Normalization. The intensifier operator is implemented in the fuzzy contrast enhancement technique. For the IFS approach, the Window Based Enhancement Scheme (WBES) and its revised version (RWBES) are applied. The RWBES gives better results compared to the WBES whereby the vague boundary of the cluster centres are reduced resulting in a smaller area of the vague boundary. The vague boundary represents the strength of the electrical potential of the foci of seizure. Next, the quality of the output image is measured via the objective measure such as mean squared error (MSE), peak-signalto- noise-ratio (PSNR), universal image quality index (UIQI), and structural similarity index measure (SSIM). In IFS, the sum of membership and non-membership is not necessarily equal to one. Thus, there exists hesitancy in deciding the degree to which an element satisfies a particular property. Moreover, the sequence of enhanced fEEG images are demonstrated by varying the value of parameter, namely �, that also influence the hesitation value π. In addition, the Sugeno type intuitionistic fuzzy generator which is used to compute the non-membership value � has been extended to the concept of fuzzy limit. Hence, by implementing the definition of fuzzy limit, different values of � will be tested in obtaining the values of integer N that will determine the value of � and hence the value of hesitation �. The relationship between membership, non-membership, and hesitation values are also demonstrated graphically.