In this paper, we present a novel cell counting method accomplished using a single-cell array fabricated on an image sensor, complementary metal oxide semiconductor sensor. The single-cell array was constructed using a microcavity array, which can trap up to 7,500 single cells on microcavities periodically arranged on a plane metallic substrate via the application of a negative pressure. The proposed method for cell counting is based on shadow imaging, which uses a light diffraction pattern generated by the microcavity array and trapped cells. Under illumination, the cell-occupied microcavities are visualized as shadow patterns in an image recorded by the complementary metal oxide semiconductor sensor due to light attenuation. The cell count is determined by enumerating the uniform shadow patterns created from one-on-one relationships with single cells trapped on the microcavities in digital format. In the experiment, all cell counting processes including entrapment of non-labeled HeLa cells from suspensions on the array and image acquisition of a wide-field-of-view of 30 mm(2) in 1/60 seconds were implemented in a single integrated device. As a result, the results from the digital cell counting had a linear relationship with those obtained from microscopic observation (r(2) = 0.99). This platform could be used at extremely low cell concentrations, i.e., 25-15,000 cells/mL. Our proposed system provides a simple and rapid miniaturized cell counting device for routine laboratory use.