Intensity inhomogeneity poses a significant obstacle in the process of image segmentation, particularly when employing region-based level set methodologies. This challenge becomes particularly pronounced when dealing with images that exhibit varying intensity values between the background and foreground objects. To mitigate this issue, a novel approach has been introduced, which leverages both global and local statistical information to enhance the accuracy of segmentation outcomes. The proposed model comprises three key components: a regularisation term, a local term, and a global term. The regularisation term facilitates the attainment of smooth and coherent segmented regions, while the local term detects areas of intensity variation by utilizing location-specific statistical data. The global term takes into account the overall intensity characteristics of the image, enabling the differentiation of items with dissimilar intensities. The effectiveness of the proposed model has been meticulously evaluated through comprehensive experimentation, which has demonstrated superior accuracy and efficacy when compared to existing models. These results serve to affirm the efficacy of the methodology and underscore its capacity to effectively address the challenges arising from non-uniform intensity distribution in image segmentation tasks.