An image sensor having a pixel array comprises periphery elements formed over a substrate, an oxide layer formed over the periphery elements, an epitaxial layer formed in an opening in the oxide layer in a pixel array area, and a plurality of photosensitive elements of the pixel array formed in the epitaxial layer. Formation of an initial metallization layer occurs after the formation of the photosensitive elements in the epitaxial layer. The photosensitive elements can thus be formed in the epitaxial layer at a higher level within an image sensor stack than that of the initial metallization layer. This advantageously allows stack height and pixel size to be reduced, and fill factor to be increased. The image sensor may be implemented in a digital camera or other type of digital imaging device.