DoktoraAçık Erişim

Novel applications of neutrosophic logic in the fields of image segmentation and pattern recognition

2018
0 görüntülenme
0 i̇ndirme
Danışman: Prof. Dr. Abdulkadir Şengür

Özet (EN)

Neutrosophy, composed of "neuter" in neutral sense from Latin and "sophia" in the sense of skill/wisdom from Greek, was introduced by Smarandache in 1980. Neutrosophy is a philosophy that brings together the knowledge of philosophy, logic, set theory, probability, and statistics. defined as an entity or event; in this case is non and is the inverse of . Neutrosophy presents a new concept, , which represents uncertainty. Furthermore, cannot be defined as neither nor . Neutrosophy is a generalization of fuzzy logic based on the assumption that t is true, i is indeterminate, and f is false. t, i, f are in a range of T, I, F and they are real values and there are no restrictions on them. Since neutrosophy has knowledge of uncertainty, many problems that cannot be solved by fuzzy logic can be solved by neutrosophic logic. In this thesis study, it is aimed to add new methods in image segmentation and pattern recognition areas using neutrosophic logic and to develop existing ones. Proposed algorithms and methods are as follows: • Kernel Neutrosophic C-Means (KNCM) clustering method for clustering datasets with noise and outliers, • A novel Neutrosophic Weighted Extreme Learning Machine (NWELM) method for classifying unbalanced data clusters, • Neutrosophic Set-based k-Nearest Neighbor (NS-k-NN) classifier, • Neutrosophic Graph Cut (NGC) is proposed for an efficient image segmentation algorithm. In order to evaluate the performance of these proposed algorithms and methods, experimental studies on real datasets, artificial datasets, and on the various images have been done. Experimental studies show that the proposed methods have a robust and efficient structure, on images and datasets which contain uncertainty such as noise and outlier data points. This thesis was supported by "The Scientific and Technological Research Center of Turkey" (TÜBİTAK), The Department of Science Fellowship and Grant Programmes (BİDEB), within the scope of "2214/A Doctorate Research Fellowship Program".

Yazar

Dr. Yaman Akbulut

Bu Yayına Nasıl Atıf Yapılır

Yaman Akbulut (Doctorate thesis). Novel applications of neutrosophic logic in the fields of image segmentation and pattern recognition, 2018, Fırat University.

Anahtar Kelimeler

Lisans

Tüm Hakları Saklıdır

Bu eser belirtilen lisans koşulları altında paylaşılmaktadır.

Fırat University tezlerinden daha fazlası