A new solution model for cardiac medical image segmentation

BACKGROUND: Calculation methods have a critical role in the precise sorting of medical images. Particle swarm optimization (PSO) is a widely used approach in the clinical centers and for other medical applications as it can disentangle optimization errors in attached spaces. In this work, a new model for image segmentation is proposed through an improved optimization algorithm. METHODS: A novel multi-objective algorithm was configured, named “multi-objective mathematical programming” (MOMP), based on the normalized normal constraint method (NNCM). In this model, the proposed algorithm was applied to evaluate the robustness of the suggested model through including the synthetic images of objects with various concavities and Gaussian noise. This model segments the individuals’ heart and the left ventricle from data sets of sequentially evaluated tomography and magnetic resonance images. To objectively and quantifiably assess the presentation of the medical image segmentations based on regions outlined by experts and the graph cut method, a set of distance and resemblance metrics were implemented. RESULTS: The numerical results obtained in experimental test cases demonstrate the validity and superiority of the proposed model through better segmentation accuracy and stability. CONCLUSIONS: The results indicated that the proposed MOMP method can outperform all traditional models in terms of segmentation accuracy and stability, and is thus appropriate for use in medical imaging.