Image segmentation is an intrinsic determinist in the performance of Computer vision applications as it directly influences the efficiency of subsequent image processing steps. Accurate identification of the region(s) of interest in an image is critical if one were to perform image analysis successfully. In this project, we discuss image segmentation with specific emphasis on scenes containing overlapping particles. This is relevant to wide range of applications, extending from analyzing grains in the food industry to rocks in the mining industry. It is the difficult task due to the lack of discontinuity between particle edges, especially they are similar textures. Determination of cell count has become more significant in health problems. If we can use a computer vision system to determine and discriminate them we will have a fast and cheap tool to decide the problem in the body and hematology’s diagnosis duration may become shorter. In this project, a segmentation algorithm based on form information is proposed for separation of touching and overlapping particles. The method integrating morphological smoothing with holes filling is employed in the pre-processing stage. Then the overlapping clusters are identified through a polygon approximation. Later the separation is implemented which consists of detecting concave points on the contours and determining separation lines. This algorithm is applied to cell cluster images. In this project, a software solution developed to determine cell count in random blood sample images. The proposed algorithm is applied to smoothly curving objects such as handwritten characters with an alteration of line approximation to polygon approximation for cell clusters. More and more image processing system is being developed to recognize overlapped regions into single ones.