Topological Neutrosophic Analysis for Uncertainty Aware Thyroid Nodule Classification in Ultrasound Imaging
Keywords:
Thyroid Nodule Classification; Neutrosophic Topology; Ultrasound Imaging; Indeterminacy Quantification; Confusion Matrix Topology; Diagnostic Confidence; Neutrosophic Neural Networks (NNN); Medical Decision Support; Explainable AI (XAI); Uncertainty Modeling.Abstract
The classification of thyroid nodules in ultrasound imaging remains clinically
challenging due to inherent ambiguities in visual interpretation, signal noise, and overlapping
morphological features. To address these limitations, this study introduces an innovative
diagnostic framework integrating Neutrosophic Set Theory with topological analysis to
quantify and interpret uncertainty in medical image classification. Leveraging
a Neutrosophic Neural Network (NNN), image features are mapped into a tripartite
representation (truth, indeterminacy, and falsity), enabling granular modeling of diagnostic
uncertainty. Further, the framework embeds classification outcomes within a neutrosophic
topological space to reveal latent relational patterns such as confidence boundaries,
ambiguity propagation, and misclassification topology that conventional metrics overlook.
Experimental validation was performed on a dataset of 1,000 thyroid ultrasound
images (Kaggle), with the proposed method achieving 92.1% accuracy, 91.4% sensitivity,
and 93.2% specificity. Crucially, topological analysis was extended to performance metrics
and confusion matrices, yielding a multidimensional assessment of classifier behavior
under uncertainty. This approach not only improves diagnostic precision but also provides
a topological lens for evaluating decision resilience, interpretability, and boundary-case
vulnerabilities. The results demonstrate that neutrosophic topology offers a novel
paradigm for explainable AI (XAI) in computer-aided diagnosis, bridging the gap between
statistical performance and clinical trust.
Downloads

Downloads
Published
Issue
Section
License
Copyright (c) 2025 Neutrosophic Sets and Systems

This work is licensed under a Creative Commons Attribution 4.0 International License.