An image processing apparatus for supporting imaging diagnosis of an eye is provided. A layer boundary of a retina region is detected from a tomographic image. A position where the detected layer boundary intersects with an upper or lower limit position of the image in the depth direction of the tomographic image is determined as a dividing position. The tomographic image is divided at the determined dividing position by a scan line in the depth direction of the tomographic image. Subsequently, whether the detection is a false detection is judged for each divided region. An average density value of the image outside the retina region according to the detected result is calculated for each divided region, and the detection in the divided region is judged to be a false detection if the average density value is equal to or greater than a predetermined threshold.