Welcome to WebmasterWorld Guest from 54.146.217.179

Message Too Old, No Replies

Google drag and drop image search

     
1:18 pm on Feb 8, 2013 (gmt 0)

Full Member from IN 

Top Contributors Of The Month

joined:June 18, 2012
posts: 294
votes: 0


Hi,

How come Google shows the exact places of an image when we drag and drop in its image search? How does it read an image even after changing the visible name of the image and changing its size from the original?

If it can understand a image, why does it ask to use image alt tags in images saying that the spiders cannot understand what an image is about?

What do I miss here? Could anybody help me?


Thanks,
1:21 pm on Feb 8, 2013 (gmt 0)

Senior Member from GB 

WebmasterWorld Senior Member andy_langton is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Jan 27, 2003
posts:2986
votes: 37


Google is visually matching the picture with others in its database. That is fine when you publish an image that is directly reproduced elsewhere (because then Google has sufficient text to guess at the image contents). But it doesn't work very well with unique images.
8:29 pm on Feb 8, 2013 (gmt 0)

Senior Member from US 

WebmasterWorld Senior Member lucy24 is a WebmasterWorld Top Contributor of All Time Top Contributors Of The Month

joined:Apr 9, 2011
posts:12702
votes: 244


Would it surprise you to hear that alt tags are not actually for search engines? They are primarily for humans who, for one reason or another, can't see the image. It does not help much to make an alt that says "picture that looks very much like the one at www.example.com/directory/images/filename.jpg".

Especially if the picture doeesn't look very much like {et cetera}.

When the drag-and-drop image search was first introduced I found some pretty hilarious results by asking for images matching a photograph of one of my rats dozing in a bakery bag. About half the results showed something with eyes in close-up, matching the middle of my picture. The other half were random scenes with strong dark/light contrast, matching the visual weight of my picture.