A study on neutrosophic $$\mathscr {T}_{\textrm{1k}}$$-semantic segmentation for iris image recognition with Gaussian and Poisson noises

In this article, we introduce an innovative methodology for image segmentation utilizing neutrosophic sets. Neutrosophic set components exhibit superior reliability in image processing due to their adeptness at managing uncertainty. The swift proliferation of neutrosophic sets research is attributed to its efficacy in addressing uncertainties in practical scenarios. Effective segmentation requires the resolution of uncertainties. This article's principal aim is to achieve multi-class segmentation through uncertainty analysis. The [Formula: see text] segmentation method pertains to type 1, encompassing truth and falsity membership functions. In this method, multiclass segmentation is possible based on the image intensity values of the neutrosophic membership functions. As a result of this research, the article proposes the finding of image segmentation through the neutrosophic set. An experimental set of biometric iris data will be the main focus of the experiment. The analysis employs real-time iris image data. The image data were sourced from the CASIA V1 iris image database. Noise was introduced into the images for analytical purposes, specifically Gaussian and Poisson noise. The evaluation metrics include the Jaccard, MIOU, precision, recall, F1 score, and accuracy. As a result of the application of this methodology, an impressive segmentation score of 85% was obtained.