An image processing device identifies the plurality of subject images which are overlapped partially, by recognizing a relative positional relationship of the plurality of subject images. The device separates into one or more blocks to segment the same area for each of the overlapped subject images the overlapping part of the images for the identified plurality of subject images. The device determines movement of an object that exists in the block for each of the blocks and generates a still image using one subject image of the overlapped plurality of subject images for blocks for which it is determined that the object has moved a specified amount or more, and using the overlapped plurality of subject images for the other blocks. As a result, the device generates a rectangular still image without blurring or doubling, because the still image is generated using one subject image of the plurality of subject images for the blocks for which it is determined that the object has moved.