The inherently varying nature of tumor shapes and image intensities make brain tumor detection very intricate.
Since several available methods and tumor detection are far from being resolved. Initially, an optimization-based
classification a new hybrid model was proposed to describe an individual use of clonal selection and particle swarm
optimization (PSO) to verify a specified MR brain image as either normal or abnormal. The methodology involves two
major stages. In the first stage, used sparse principal component analysis (SPCA) to reduce feature space, and selected the
important features. The second phase two hybrid optimization-based negative selection models were developed to
investigate the integration of clonal selection technique with PSO from the perspective of classification and detection to
optimize the parameters C and 𝜎. Fivefold cross-validation was utilized to avoid overfitting and to ensure a robust
classification. Although clonal negative selection classification algorithm (CNSCA), has the best performance. The
proposed method achieved 99.10% classification accuracy. The admirable features of the outcomes submit that the
suggested methods may institute a basis for reliable MRI brain tumor diagnosis and treatments. A comparison with other
techniques showed the competitiveness of the proposed methods.
الاستاذ المساعد الدكتورة اسماء شاكر عاشور | Assist. Prof. Dr. Asmaa Shaker Ashoor | 18426