A method to form a stacked CMOS image sensor includes forming a signal processing layer including a plurality of discrete signal processing circuit, an image sensor layer including a plurality of discrete image sensing units, and an intermediate capacitor layer including a dielectric layer and a plurality of capacitors. Each capacitor includes a first electrode, a V-shaped or U-shaped first electrode material layer electrically connecting to the first electrode, a second electrode material layer on the first electrode material layer having the dielectric layer there-between, and a second electrode electrically connecting to the second electrode material layer. The method further includes bonding the signal processing layer to the intermediate capacitor layer with each second electrode electrically connected to a signal processing circuit, and bonding the image sensor layer to the intermediate capacitor layer with each first electrode electrically connected to an image sensing unit.