Symbiotic AI for Medical Imaging: Anomaly Detection and ExplanationGenerationUsing KnowledgeGraphs
| dc.contributor.author | •Fatma Zohra BOUSSOUFA | |
| dc.contributor.author | Dr.Youcef BEZZA | |
| dc.date.accessioned | 2026-07-09T09:16:18Z | |
| dc.date.available | 2026-07-09T09:16:18Z | |
| dc.date.issued | 2026-07-09 | |
| dc.description | Keywords:Brain Tumor Detection; MRI;CNN; OWL Ontology;SymbioticAI. | |
| dc.description.abstract | Medical imaging plays a crucial role in disease diagnosis, monitoring, and treatment planning. AlthoughConvolutionalNeuralNetworks(CNNs)achievehighaccuracyinanalyzingMRI,CT,and X-ray images, their black-box nature limits clinical trust and interpretability.Moreover, reliable diagnosis requires integrating visual evidence with structured medical knowledge.To address this challenge, this project proposes a Symbiotic Artificial Intelligence framework that combines a fine-tuned VGG16 CNN with a brain tumor ontology developed in OWL and enhanced by SWRL rules covering ten categories of medical inference.Using Owlready2 and RDFLib, the system performs ontology-basedreasoningtogenerateaninterpretablediagnosticprofileforeachMRIimage, including diagnosisstatus,malignancyclassification,expectedsymptoms,recommendedtreatment,confidence level,andCNN–ontologyagreement.Theprojectwascompletedoversixmonths,includingliterature review, implementation, experimentation, and report writing.Trained and evaluated on 10,800 brain MRI images (5,400 tumor and 5,400 non-tumor), the proposed framework achieved a test accuracy between98%and99%, demonstratingthefeasibilityandclinicalrelevanceofintegratingdeeplearning with knowledge-based reasoning for explainable brain tumor detection. | |
| dc.identifier.uri | http://dspace.cuillizi.dz/handle/123456789/537 | |
| dc.publisher | المركز الجامعي إليزي - معهد العلوم، قسم الإعلام الآلي | |
| dc.title | Symbiotic AI for Medical Imaging: Anomaly Detection and ExplanationGenerationUsing KnowledgeGraphs | |
| dc.type | Other |