An Optimal Similarity Neutrosophic Model Based on Distance Measuring to Improving Content-based Image Retrieval

This paper deals with images using the theory of neutrosophic, which the idea of working, on set about the degree of truth, indeterminacy, and falsity. Which helped to discover the hidden features of the images that were segmented by using neutrosophic image processing into objects and then extracting the features into the three truth, indeterminacy, and falsity levels of the image and combining these features to extract the original image features.
The proposed similarity model namely weighted Hamming distance measure that based on the single-value neutrosophic set was used to retrieve images from the database, by matching with the query image that extracted its feature in the same way.
The results showed that the proposed system is highly efficient in retrieving images compared to different distance measures such as Euclidian, Manhattan, and Minkowski. Finally, A novel similarity model used to match the neutrosophic image features for CBIRs. In the proposed system, an image is segmented into objects, edges, and backgrounds by using neutrosophic image processing.