The Edge of Image Uncertainty

About the Project

Active Contour Models (ACMs) are widely used for image segmentation but struggle with uneven image intensity and selecting specific objects. This innovation proposes SFGL, a selective fuzzy region-based model combining global and local fitting energies with marker points to improve segmentation accuracy and efficiency. Unlike existing models, SFGL accurately identifies selected objects while reducing computational complexity. Experiments involving 20 images across five categories showed that SFGL outperformed existing models, achieving average Jaccard (0.817), Dice (0.895), and Accuracy (0.985) scores. With an average processing time of 1.80 seconds, SFGL provides a faster and more accurate approach to selective image segmentation.