Performing image processing is disclosed, including: causing an original image to be displayed in a first editable object of a page, wherein the first editable object is generated based at least in part on web page data from a server; receiving an input track corresponding to the original image of the first editable object of the page; selecting at least a portion of the original image relative to the input track to use as a result image; and causing the result image to be displayed in a second editable object, wherein the second editable object is generated based at least in part on the web page data.