Traditionally, the population of human embryonic stem cell (hESC) culture is estimated through haemacytometer counts, which include harvesting the cells and manually analyzing a fraction of an entire population. Obviously, through this highly invasive method, it is not possible to preserve any spatial information on the cell population. The goal of this study is to identify a fast and consistent method for in situ automated hESC population estimation to quantitatively estimate the cell growth. Therefore, cell cultures were fixed, stained, and their nuclei imaged through high-resolution microscopy, and the images were processed with different image analysis techniques. The proposed method first identifies signal and background by computing an image specific threshold for image segmentation. By applying a morphological operator (watershed), we split most physically overlapping nuclei, leading to a pixel area distribution of isolated signal areas on the image. On the basis of this distribution, we derive a nucleus area model, describing the distribution of the area of cell debris, single nuclei, and small groups of connected nuclei. Through the model, we can give a quantitative estimation of the population. The focus of this study is on low-density human embryonic stem cell populations; hence cultures were measured at days 2-3 after seeding. Compared with manual cell counts, the automatic method achieved higher accuracy with <6% error.