| 摘要 |
An image processing apparatus that is capable of improving the alignment accuracy by detecting a motion vector accurately even if there is a low contrast region or a repeating pattern region. A division unit divides each of inputted images into blocks of a predetermined size. A first determination unit determines whether contrast in a block is less than a predetermined contrast for each of the blocks. A size changing unit changes the size of the block concerned when the first determination unit determines that the block concerned is low contrast. A motion vector detection unit detects a motion vector by comparing images in each pair of corresponding blocks between the images. |
