Forum Moderators: open
The page this image is on also includes the keywords in other parts of the text, plus the images are organized in tables. The description for this particular image is in a cell directly below the image and uses the text "Great keyword!"
Google also uses sophisticated algorithms to remove duplicates and ensure that the highest quality images are presented first in your results
Which parameters would qualify for "highest quality images"?
with more than 390 million images indexed
At an actual Google index size of 3 billion, thats probably a lot of text only pages and also a lot of duplicate images. Either that, or their Imagebot is way more selective.
As I just said, I try to keep my images out of the Google index, so hearing that there was an update going on I decided to do a check. Here's something that I don't think should be happening: on one page I have two gifs that are included as <img>s from another site. That is, on mypage.org I include two <img>s with src=othersite.org/alpha.gif and src=othersite.org/beta.gif. If you search Google's image index for "mysite" it pulls up both of those gifs, and the address it lists under them is "mysite.org" as though the gifs were really on my site instead of being transquoted (as Ted Nelson would say).
I do understand the logic of returning those images when I search for the word "mysite", but I don't find it intuitively correct to list mysite.org as the address under them; "quoted on mysite.org" would seem closer to the truth, and Google can certainly tell from the URL where the image is coming from. I suppose this is a minor interface point, but it really struck me as odd.