Hi Wilhelmina, you might want to take a look at the work of Myriam Ben Saad: http://www-poleia.lip6.fr/~bensaadm/. She has a few papers describing extending a vision-based page segmentation algorithm (VIPS) to support webpage change analysis (Vi-DIFF). Here's one such paper from 2010: http://www-poleia.lip6.fr/~bensaadm/DEXA2010.pdf. The algorithm detects both visual changes and changes in the DOM.
Unfortunately, I haven't yet seen any links to tools or production code using the approach.
~Nicholas
--
https://twitter.com/nullhandle
|