Favourable traits It may be difficult to partition real-time photos because of their frequent uniformity. The majority of segmentation methods for photos are based on regions and typically rely on the uniformity of the picture intensity in the regions of interest (ROI). But because the ROI intensity is constant, these algorithms frequently produce inaccurate segmentation results. In this, a novel technique for segmenting areas of a picture that can handle homogeneities in intensity is proposed. first the intensity-based picture model. In order to infer a local intensity clustering characteristic of the image intensities in homogeneities, we develop a local clustering criterion function for the image intensities adjacent to each point. A global criterion of picture division is applied after integrating this local clustering criteria function with respect to the neighbourhood centre. This criterion defines energy in terms of level set functions that represent a partition of the image's domain and a bias field that, in a level set formulation, accounts for the intensity in the homogeneity of the image. Therefore, our method can segment the image and estimate the bias field simultaneously by using the level set approach and lowering this energy. To account for intensity in homogeneity, the estimated bias field can then be used for additive bias correction. Our method has been tried on real and artificial images in many modalities, and it works magnificently even when the intensity is homogeneous. Experiments revealed that our method is quicker, more accurate, and more resistant to beginning smooth piecewise models than the widely used one. Our technique has been used to segment photos and correct bias with good results.