Skip to main content

Neutrosophic image segmentation with Dice Coefficients


Sudan Jha, Le Hoang Son, Raghvendra Kumar, Ishaani Priyadarshini, Florentin Smarandache, Hoang Viet Long*

Source title: 
Measurement, 134: 762-772, 2019 (ISI)
Academic year of acceptance: 

This paper explores various properties of Neutrosophic sets (NS) and proposes a novel idea on Image Segmentation using NS. A theoretical Neutrosophic model is proposed to reduce uncertainty from missing data. Besides, we also tackle the problem of image segmentation with fewer assumptions. Min-Max Normalization is used to reduce any uncertain noise in an image due to a number of factors during image capturing. Next, we apply activation functions to resolve the non-linearity in the image followed by the computed membership functions. These sets are then transformed and compared with others to find similarities and dissimilarities. Neutrosophic Sets and Dice’s Coefficients are fused to ensure proper evaluation of uncertainty of the missing data and their indeterminacy for image segmentation. The proposed method is experimentally validated.