Just hours after GazoPa announced its similar image public beta test to the public, Google announced that it was releasing its similar image search out of Labs. Now, anyone could have experimented with the similar image search while it was in Labs, but you can only run one Labs experiment at a time. And a new experiment, Social Search, was added to Labs yesterday.
We're used to seeing Google release updates, products, major news, etc. in reaction to Microsoft or Yahoo! but saying "Me, too" to a smaller site just seems weird. Not everything that Google has done in this light has been ready (cough, Google Squared, cough). So, how does similar images fare?
To try it out, go to Google Image search. I used "pumpkin" to test it out since that's the keyword I used with GazoPa earlier. Thankfully, Google returned images of pumpkins. GazoPa only returned one or two even with keyword search.
The red circle in the graphic above highlights the "Find Similar Images" link, which I hope is self-explanatory. Sure enough, images of pumpkin similar to the one I clicked on appear in the results.
Unfortunately for GazoPa, Google is already light years ahead in similar image search. That doesn't mean GazoPa could never make inroads, they just have their work cut out for them.
Posted by Nathania Johnson at 12:19 AM | Permalink | Comments (3)
Back in March, Google updated its mobile image search but just for iPhone and Android devices. Now, Google says just about every feature phone in 38 languages is getting the update. The update includes larger thumbnail images as well as search filters.
Something to know about the results is that when you click on an image, it takes you to a detail page on Google image search instead of to the page where the image is hosted.
To use Google image search on your phone, simply browse to Google.com and select "Images." You can also just click "Images" after you conduct the search.
Posted by Nathania Johnson at 6:05 PM | Permalink | Comments (2)
Remember back in May when Google launched Search Options as part of its Searchology day? Well, now they're rolling it out to Image Search as well.
If you don't remember, Search Options provides a panel on the left side of the results page which helps you filter the results. Search Options originally launched just on the main Google web page.
As pretty much all of Google launches, this one is being rolled out. I wasn't yet able to access it.
How about you? Have you checked out Search Options on Google Images yet? Tell us what you think of it in the comments section below.
Posted by Nathania Johnson at 11:45 PM | Permalink | Comments (0)
When you're traipsing around the earth via Google Maps, you now have a new option to get a little closer. Google is incorporating photographs from Panoramio, a travel photo sharing site, to assist with Street View zoom.
Let's say, for example, that you're checking out the Eiffel Tower in Street View from Avenue de la Motte Picquet. To access the new feature, click the user photos box in the top right corner (yes, this was added back in February, hold your horses, I'm getting to the new stuff.)
Once you've done that, check out the shapes on the image of the Eiffel Tower. (Or click one of the images in the carousel.)
Mouseover one of the polygons to select an image to view. Then, double-click to actually view it.
The new image shows a zoomed in look at the area you selected:
Then you can select yet another polygon on the new image:
And see yet another image that's zoomed in a little more:
You can keep doing this until you're as zoomed in as Panoramio photos provide.
What do you think of this new Google Maps zoom feature? Let us know by zooming below to leave us your comment.
Posted by Nathania Johnson at 12:40 AM | Permalink | Comments (2)
Google has added a Creative Commons filter to Image Search. In order to use the filter, you'll need to use the Advanced Search option in Google Image Search.
Once on the Advanced Search page, look for the "Usage Rights" option, which is the second from the bottom. You won't see "Creative Commons" listed as an option. Instead, you'll see options for:
Google's not the first to do this. Of course, Flickr has had Creative Commons search for a long time. Yahoo! added a Creative Commons filter to its Image Search last May.
For its part, Google added Creative Commons filtering options to Custom Search plus YouTube began offering Creative Commons licensing this past February. However, YouTube doesn't yet offer Creative Commons search in its Advanced Search yet.
Posted by Nathania Johnson at 12:17 AM | Permalink | Comments (1)
Google has updated the universal search results for location-based queries. Now, next to the map included at the top of the search results are images.
It doesn't work for all location-based searches, however. A search for Kota, India shows image results but they are further down and no map is included.
A search for Nassau, Bahamas returns a map but not images:
It's not a problem with international searches altogether because a search for Brasilia, Brazil returns the new results:
Still, if you're invested in travel or local search, you'll want to spend some time optimizing for image search, since they're now a big part of Google location-based results.
Posted by Nathania Johnson at 12:10 PM | Permalink | Comments (0)
Google employees get 20% of their time to work on non-assigned projects. Sometimes when they do, the result is some pretty cool stuff that Google adds to its products.
This time, Google has announced that two 20% projects have made their way into Google Labs (aka Google's experimental playground). The two projects are "Similar Images" and "Google News Timeline."
Last week, we brought you word that Google News had introduced a "Timeline of Articles." This new timeline that's launching in Labs is different.
The Labs timeline looks like a calendar with a bunch of news stories. An on-site scrollbar helps users find more stories for a given day:
Meanwhile, Similar Images does pretty much what the name implies. Start off by searching for an image. In this case, I typed in the name of one of my favorite shows on TV, Chuck (starring Zachary Levi, don't cancel it, NBC, don't!):
In the results, I look for an image I'm interested in and click the blue link underneath that says "Similar Images"
As you can see, the results are really good!
You may remember that recently the official Google Image search recently added a color filter. It works quite good as well. It's nice to see these advances in image search from the Google team.
Posted by Nathania Johnson at 12:13 AM | Permalink | Comments (1)
Google has added a handy color filter to its image search. Simply conduct the search and then look for the color box to filter down the search.
Interestingly, you get different results if you actually type the color in with your search. The screenshots below show:
1. a general search for "bike" 2. a general search for "bike" with color filter set to "green" 3. a search for "green bike" with color filter set to "Show all colors"
(Also, color filter didn't show up for a search for "ice cream")
Related Reading: Google Image Search Adds Search By Content Type Google Launches New Image Search for iPhone, iPod Touch, and Android
Posted by Nathania Johnson at 2:53 PM | Permalink | Comments (2)
iPhone, iPod Touch, and Android users in the US, UK and Japan are getting a brand spanking new image search from Google. Now image search results on these platforms return 20 results on a single page.
From the results, users can view a larger thumbnail image, visit the site where the image is hosted, or view the image alone in its full size. The new version also includes a filter where users can narrow their results to display people's faces, clip art, line drawings, or photo content. This type of filter was added to the regular Google Image Search last December.
To use the search, browse to Google.com on the iPhone, iPod Touch or Android browser and then select "Images" before searching.
Posted by Nathania Johnson at 9:01 AM | Permalink | Comments (0)
If you've ever conducted a search on Google Image Search and wished you could separate the photos from the clip art, then you're getting your wish. Now you can use a drop down menu to sort through the specific type of image you're on the hunt for.
Here's a screenshot:
Related Reading: Google Image Search Hosts LIFE Image Archive Google Adding Ads Across Its Web Properties Thanksgiving Turkey, Image Search, and Missed Opportunities
Posted by Nathania Johnson at 12:13 PM | Permalink | Comments (0)
Google has announced that it has made LIFE Magazine Images available as an archive. The images from the iconic magazine are also available through Image Search. You can view the collection here.
So far, 20% of the total collection is online. The entire archive consists of 10 million photographs. Who are the lucky interns who get to digitize negatives, slides, glass plates, and prints?
Images can also be purchased from LIFE. You'll see a link for that on the right hand side when viewing an image. But it won't be cheap.
A framed 1938 photograph of Katharine Hepburn on the set of the play "The Philadelphia Story" (yes, the play not the movie) will run you $79.99.
Related Reading: Google Adding Ads Across Its Web Properties Google Will Bank on VisualRank - PageRank for Images
Posted by Nathania Johnson at 11:02 AM | Permalink | Comments (1)
Like search ads, PageRank, and blended search, Google's newly launched VisualRank image search technology has the potential to change the way SEO is done. In today's Searching for Meaning column, "Visually Impaired Search," Kevin Ryan explains that while the game hasn't changed yet, it will very soon.
Posted by Kevin Newcomb at 12:00 AM | Permalink
Thanksgiving is a major holiday for my family. With the first changes of color in the foliage, I begin the counting the weeks until the family gathers at my cousin's for this event. So, with Thanksgiving less than a week away, I decided to share my mounting enthusiasm with one of my siblings by sending a note with a picture of what I hoped would be a gorgeous roasted turkey. I fully expected that finding such an image would be as simple as firing up my browser turning to either Google or Yahoo image search and selecting a choice roasted turkey. To my surprise this was not to be.
Here is what I found when I went turkey tracking. An image search for “turkey” at both Google and Yahoo! yielded differing, but unsatisfactory results. Google's included live turkeys, some turkey humor, maps and images of the country Turkey, but no succulent roast turkeys on page 1. On page 1 of the image search results Yahoo! had a number of live gobblers, no maps, some images of Turkey and Turkish people, and lone image of a roast turkey – not quite what I had in mind. Both search engines suggested that I narrow my search to include “Thanksgiving turkey.” I took the suggestion, and the results were somewhat better. Google offered up more turkey clip art and a few images of roasted turkeys. Yahoo! had similar results.
As I reviewed the results, I got to thinking – a dangerous activity – where are the large purveyors of turkeys? Why are there no images from Butterball or Perdue, two large poultry producers or even the food magazines that every year have many gorgeous images of roasted turkeys within their covers? I thought initially that my mistake was not searching correctly so I gave image search another try using “butterball turkey” and “perdue turkey.”
Now we're cooking, I thought. The results in Yahoo! for “butterball turkey” were short on images of cooked turkeys, long on displays of the logo. There were also disturbing images of maimed and dead (not in grocery settings) turkeys and images chronicling a turkey “offal to oil” initiative. With Yahoo! it wasn't until the second page of results that a search for “butterball” turkey yielded a beautiful cooked turkey.
A search for “perdue turkey netted on Google netted a wide variety of images including some appetizing pictures of turkey sandwiches. On Yahoo! the results were startling, for I could have filled a grocery bag with raw turkey parts in neat packages, but no beauty shots of roast turkey.
Still not satisfied, I made one last try using “roast turkey” as my image search term. Google offered up lots of images of roasted turkeys, but curiously not a single one on the first page was from a turkey producer or packer. On Yahoo! there were lots of roasted turkeys some with a more homey touch since they were from Flickr, but again not a single one on the first page from a turkey producer.
Just to get a rounded picture I searched for “thanksgiving turkey” in the main “web results” pages for Yahoo! and Google. Lo and behold, Google offered a lovely cooked turkey, from a family advice site, not a turkey producer. Yahoo! had turkey clip art but no roasted turkey.
Is this a missed opportunity? Surely, I am not the only person searching for a beauty shot of a roasted turkey? Were my searches that defective? My persistence at least should have been more easily rewarded. I even took the help that the search engines offered. I'm left with the conclusion that image search has been neglected. A visit to the turkey sites reveals very sophisticated marketing programs for assisting the consumer in preparing the ritual bird. There are hotlines, chat rooms and podcasts, but alas, lowly image search appears to be overlooked. With the advent of universal search, businesses must be found in all the right places. (Note: this is the title of the workshop that I will be giving at SES Chicago along with Greg Jarboe.)
Oh! When all was said and done, I decided not to send a beauty shot, but instead to go with the image of the roasted turkey in a bikini – an effect that can be achieved with creative use of foil and the tanning effects of oven roasting.
Posted by Amanda Watlington at 4:25 PM | Permalink
Google is celebrating its 9th birthday today. With little fanfare, except for the logo change, the search engine has obviously had more pressing things on its mind - like dealing with the Senate Subcommittee on Antitrust, Competition Policy and Consumer Rights Committee on the Judiciary.
Interestingly though, if you google Google, they have the 9th birthday announcement in their SERP description. I wonder if Larry and Sergey will go to their favorite Burger King store and reminisce about the old days?
Posted by Frank Watson at 2:19 PM | Permalink
Google executives discussed the possibility of bundling image or video ads into Google Universal Search, during the CitiGroup Technology Conference in New York this past Thursday.
As reported by Tameka Kee at MediaPost, Google's group business product manager Nicholas Fox says that Google has had internal discussions on how to incorporate visual advertisements in search engine results pages (SERPs) that best match search query relevancy.
According to Sundar Pinchai, Google's director of product managment, Images and video have the potential to be more relevant than simple text ads in at least some situations, but cautions that "the images and video ads you see today on content networks are not what will work." Incorporating these type of ads in SERPs could lead to ad blindness and hurt business in the long-term.
While Google's comments suggest that multimedia advertising will eventually be established to some degree in universal search results, they are not something for advertisers to include in their plans for this year. Google says that any approach they undertake with incorporating multimedia ads will be slow, cautionary, and incremental.
Currently, Google allows advertisers to run both video and image ads in their Contextual network, which is separate from search. Those ads, along with text ads, have been claimed by search advertisers of having lower content relevancy than with SERPs. While Google has been making some appeasements with adjustment ad pricing and network placement control advertisers in the contextual program, advertisers already feel their multimedia creatives would be much better suited directly to search.
Will all things considered, Google is correct in determining that ad relevancy for multimedia will have to be higher when showing up on its own search pages than how they currently have them on 3rd party sites.
Posted by Grant Crowell at 12:36 PM | Permalink | Comments (0)
Google Image Search is in the process of adding face-recognition technology, as well as new features to filter out images from news stories, according to Google Blogoscoped. The features, which so far are available only by appending a variable to the end of a search URL, are apparently based on technology from Neven Vision, a company that Google acquired last August.
For a search of faces, the filter is &imgtype=face. For example, a search for images related to Search Engine Strategies would return several kinds of images, including people, buildings, presentations and company logos. With &imgtype=face at the end of the URL, the same search shows only people.
The news image search filter is &imgtype=news. Compare a search for whales without the filter to one with the filter. The news image search returns mainly photos related to recent news stories about whales, while the unfiltered search includes many more kinds of whale images.
So far, Google has declined to comment on the changes, or to say if any other search qualifiers are expected to be launched for image searches.
Posted by Kevin Newcomb at 3:19 PM | Permalink
Abbey Klaassen of Advertising Age reports that Ad Age's Digital Fact Pack is available to download for free. Among the highlights in the second annual Digital Fact Pack are these gems:
-- The top 10 online properties took in 99% of 2006 gross online ad revenue. Or, as Klaassen puts it, "the Long Tail of the web has a big, fat head."
-- MySpace and Facebook continue to defy gravity, growing 72.5% and 59.2% (Feb. '07 vs. Feb. '06), respectively.
-- The top US Search engines are Google (51.83%), Yahoo Search (15.94%), MSN Search (9.13%), Google Image Search (6.02%), and Ask.com (2.15%), according to Hitwise. That's right, Google Image Search has a higher market share than Ask.com!
Posted by Greg Jarboe at 11:42 AM | Permalink
Users of Google Image Search will find a changed, redesign results page. The search results no longer present the image, file name, dimensions, size and URL. Instead, there is now just the image and a brief contextual snippet. Users are forced to hover their mouse over the image to retrieve the size, file type and domain. The new design also offers the user an opportunity to narrow the search for large, medium or small images. I'm personally not sure how I will like this design. Although it will be nice to filter by the size of image desired, this new cleaner interface makes it more mouse-work to get at the detail information.
Posted by Amanda Watlington at 8:59 AM | Permalink
Kate Goodloe of The Wall Street Journal Online reports that an online show, called "News at Seven," is using "an automated computer program to comb online news outlets for major stories of the day and to pair them with video and still photos culled from sites like Google Images and YouTube. The newscast is delivered by an avatar."
Still in beta, News at Seven is a project of two computer-science graduate students and a professor at Northwestern University's Intelligent Information Laboratory, and was funded with a grant from the National Science Foundation.
Posted by Greg Jarboe at 10:26 AM | Permalink
Google Blogoscoped spotted the Google Image Labeler game, designed to help Google improve its image search results through tagging. It feels like a catch-up game with human-powered efforts that Yahoo is embracing via Flickr -- plus it also looks pretty influenced by the work of Luis von Ahn and his ESP Game.
The game pits you against someone else. If you see a picture of a car, and you both label it car, you can proceed to the next image. You continue until your time has run out. Here is an image of Horcrux and Barry Schwartz from our blog (rustybrick) scoring 300 points for matching tags on three images.
Image search has been tough for search engines. They can't easily tell what an image is about, since there's no good way to "see" the images and categorize them. Some technologies to recognize faces, colors, shapes, objects and other things are improving. Still, it's hard with an image of someone like Martha Stewart. Is she a woman, celebrity, criminal or just Martha Stewart? Or all of these? How do you know which one or ensure that all of them are applied.
Yahoo's Bradley Horowitz is probably one of the most famous converts from turning to human power over computer power. He's been cited many times as having originally sought a technological solution to understanding what's in video and image data, then moved to embrace people power. Here's one example of that from a Wired article last year:
Horowitz's favorite project is incorporating people-powered metadata systems from two other Yahoo! properties: the recommendation technology from Yahoo! Music and the tagging features from Flickr, the photoblogging company Yahoo! acquired this spring. Google's original stroke of genius was figuring out how to piggyback on human judgment by following the links people make between Web sites. Horowitz is borrowing functionality from two Yahoo! properties to develop something similar for video.
We've just seen Yahoo make more of a commitment to using that human power when it started inserting Flickr results, rather than Yahoo Image results (which are computer sorted), into regular web searches last week.
Google, of course, has no Flickr to use. Enter the game. It's designed to get lots of people to quickly label images because they want to have fun. If that concept sounds familiar, it's because that's exactly the method behind the ESP Game, created by Carnegie Mellon professor Luis von Ahn.
I first learned of Ahn's work through a 2003 Associated Press article, Researchers Hope to Improve Web Searches. His ESP Game came later. The Google system looks like a copy of it, perhaps with his cooperation. Perhaps he's even there now -- I'm checking. We know he was there just last month, because here's a video of his lecture on classifying images (well worth watching). That Martha Stewart example above? That came from his video. I've also embedded it below:
In his video, Ahn thinks that in two months, all images on Google Images could be labeled. It's not hard to believe after hearing that, Google jumped to try this.
Postscript: I heard back from Luis von Ahn, who sent me this:
Yes, Image Labeler is based on my ESP Game, which Google licensed. I'm not employed by Google, however, since I'm a full-time faculty member at Carnegie Mellon.
Posted by Danny Sullivan at 9:43 AM | Permalink
Amit Agarwal has a nice write-up on how to increase your chances of listing your images high in Google Image Search. The tips include;
+ Use a descriptive and keyword rich file name + Use seven to eight keywords in your title and alt text + Place a short one line description of the image directly below the image + Wrap the content around the image + Try to put the image closer to the top of the page
Those are just a sampling of the tips to get listed high in Google Image Search.
Posted by Barry Schwartz at 8:55 AM | Permalink
Philipp over at Google Blogoscoped has the story about text snippets seen below each image found via a Google image search (as seen below image search thumbnail). This new feature was noticed and reported on yesterday and is now live to SOME Google Image Search users. After first thinking that it was live to all Google image search users, Phillip (one of the hardest working people in Googledom) has just updated his post noting that this "experiment?" is still not visible to everyone.
It looks like (no pun intended) this might be yet another Google UI test? When I ran an image search a few minutes ago the text snippets were not visible.
Postscript: Google has confirmed with SEW Blog that this is another UI experiment.
Postscript 2: I'm now seeing the results page test.
Posted by Gary Price at 1:49 PM | Permalink
Over the weekend I went searching for Hurricane Katrina imagery in the Google Images and Yahoo Images databases. It's been two weeks since the storm hit the now devastated gulf coast and I wanted to determine what a typical user might find when entering the simple but descriptive query [Hurricane Katrina] in the image search box at Google and Yahoo.
At Google Images only 36 images (I'm not kidding, are seen on the results pages (screen cap here) and the relevancy of a few of these images are questionable (they have nothing to do with the storm and the devastatlon it caused).
Btw, I ran the search several times over the weekend to make sure a technical error was causing so few images to appear.
Many of the images that Google does offer are charts and maps of the storm before it slammed into the gulf coast. If you're looking for imagery that illustrates the destruction Katrina caused, you'll find very few. I was surprised that with the tens of thousands of images found on the open web, I would have found more. Perhaps this exercise also gives us some idea about how often Google Images is updated?
How did Yahoo Images do with the Hurricane Katrina query?
MUCH better than Google.
The total estimated number of images for the [Hurricane Katrina] query at Yahoo Images is over 6,100 but that number means little since I was only able to view 531. That said, most -- but not all -- of these images have a direct relationship to the storm and its aftermath. You'll find images from various sources culled from the open web along with imagery available from Yahoo News.
One thing I was unable to find via in either Google Images or Yahoo Images was material from the FEMA Photo Library, U.S. Dept of Defense Image Collection, and other sources inclding Orbimage and DigitalGlobe. Of course, this once again illustrates why knowledge of specialized/fcoused databases is so important.
Posted by Gary Price at 10:59 AM | Permalink
Tara Gives Thumbs Up To Yahoo Images; Yotophoto For Copyright Free ImagesContemplating Yahoo Images and Google Images is a short item from Tara over at ResearchBuzz finding that in a search for [rss buttons], less is more. Yahoo had fewer results than Google but was more on target with what she wanted. Meanwhile, her Search Engine of Free, GNU FDL, Creative Commons Images looks at the new Yotophoto image search engine designed to give you images you can use without copyright worries.
Posted by Danny Sullivan at 9:45 AM | Permalink
If you asked me yesterday after Yahoo's total index size announcement (web, images, audio databases) what I would be posting today I would have said that Google would post a new total size to one or more of their databases. I would have been correct.
Google has now posted a new total image size count on the Google Images home page.
The new total listed is 2,187,212,422 up from 1,305,093,600. Yes, that's nearly double. In yesterday's announcement Yahoo said their image database currently contains 1.6 billion images.
Remember, Yahoo and Google's numbers are just claims and they're mire about bragging rights and for keeping the buzz going especially in the non-search community. Many other factors (relevance, freshness, etc.) are what really matter.
Posted by Gary Price at 8:03 PM | Permalink
Amazon Sued for Copyright Infringement from the Associated Press covers how Amazon is being sued by Perfect 10, an adult magazine and web site, over Perfect 10 images apparently appearing in Amazon-owned A9 image search results. Those results are powered by Google, which we've previously blogged about being sued by Perfect 10 back in November.
Posted by Gary Price at 9:22 AM | Permalink
Although they're just rough estimates that are primarily best used for marketing purposes (aka bragging rights), I guess it's worth noting that Google has just made an increase to the public number they post for total number of images searchable at Google Images. The new total number is 1,305,093,600 images an increase of 117,463,600 from 1,187,630,000.
I would have thought that when Google posted an increase to the total images number it would have been greater than 1.5 billion. Why? This is the number that Yahoo used in an announcement about 4 months ago.
Posted by Gary Price at 6:58 PM | Permalink
Tim Bray tests whether Google, Yahoo and MSN image search services can find images that make use of keywords only within the the ALT and TITLE attributes of an image tag. Google is the only one that can. The Cherry-Tomato Challenge and A Cherry-Tomato Winner tells the story.
Posted by Danny Sullivan at 12:48 PM | Permalink
Google has announced its Google Images database now includes 1.1 billion images. In addition, it confirms what some have been seeing before, that images will now be displayed inline with regular web search results, depending on the query. For example, a query on sunsets or mountains is supposed to bring back images (for me, they don't yet do this).
Google Images suffered embarrassment last year when it was revealed that the image database hadn't been refreshed for some time. Stale & Split Image Databases Fuel Google Conspiracy Theories and More On Google Images & Google News Images cover this more.
Ask Jeeves has long done inline display of images in the way Google is now doing, as Gary covers in this previous post. Yahoo doesn't do inline display but will prompt with a text link at the top of results to try an image search as this example shows: pictures of sunsets. Yahoo also increased its own image database last October, and Gary sings its praises in his Yahoo Announces Size Increase to Image Database.
Also from Gary, a collection of other image search databases can be found here: Searching For Images. In addition, Ditto and Picsearch are image search services worth checking out.
Posted by Danny Sullivan at 3:50 PM | Permalink
Google Images In OneBox Display Further Supplant Main ResultsWe've had sporadic reports of Google inserting image matches into web results, and Barbara Coll just sent an example for "united states" that finally works for me and hopefully for you as well. Also try this one for "united kingdom" or try some other countries.
More important, in the first "united states" example, notice that above the image results insertion, there's also a book results listing.
Google's long done what it calls OneBox insertions like this. But to date, I've never seen more than one insertion. You'd be shown product results, like this for dvd players, or news results like this for iraqi elections but never more than one OneBox insertion with one major exception: Google Desktop.
Those running Google Desktop might see first Google Desktop results as a OneBox insertion, then a Google specialty search display, like product search. Now with the image search testing, that example gives me an unprecedented three OneBox insertions.
The insertions also mean that the "main" or "organic" listings move further down the page. This is something I touched on in my SES Chicago keynote last December -- anyone not exploring vertical search had better do so. I illustrated in before and after fashion how such insertions by Google and others are pushing down the "main" results.
For searchers, this is generally good news. Quite often, they should be going to some of the specialized databases out there. For search marketers, this is also good news as long as you are paying attention to the vertical searches being shown. Failure to do that will mean failure to succeed, in the long term. For more on that see:
Posted by Danny Sullivan at 8:26 AM | Permalink
Philipp over at Google Blogoscoped points out that Google is testing placing few thumbnail images from their image database at the top of web search results pages. I tried a few searches but wasn't able to see any examples. Luckily, the GB post also includes a screen capture.
Btw, embedding a few thumbnail images directly onto a page of web results is not a new idea. Ask Jeeves has done this (if the query indicates the searcher is looking for images) for almost two years. Here are a couple of examples from AJ: + images elephants + pictures golden gate bridge
Posted by Gary Price at 5:56 PM | Permalink
Via John Battelle, news of an AP reporter locating new photos through Google of possible abuse of Iraqi prisoners by US soldiers that might predate those that came out earlier this year.
Navy Probes New Iraq Prisoner Photos from the AP details explains how the reporter came across the pictures after performing a Google search to research a story about new allegations of abuse.
The story doesn't provide details about exactly how the search was performed. It sounds like an ordinary Google web search was done, which lead to an online photo album hosted by picture sharing site Smugmug.
That page, which is not identified by URL but shown in this version of the AP story, is said by the AP to now be password protected. Google's cached version of it, again not identified, is said to remain in Google.
The photos might also have been found via Google Images, but I suspect not. The reason is because the AP story mentions that the "archived" version of the page remains in Google. If it were a Google Images search that had been involved, I'd have expected the reference to be about the "snapshots" of the photos themselves remaining.
Some will no doubt recall the issue of Google Images not finding pictures of Abu Ghraib prison abuse, which Google has said is due to its image database being several months out-of-date. These new photos, in contrast, might have been in Google Images given they appear to be from last year.
So, if you read about Google Images finding "new" photos of abuse when it couldn't find these older ones, keep in mind:
FYI, I'm still investigating Google's explanation about why the Abu Ghraib photos that at least one person says he saw now don't appear in Google Images. My More On Google Images & Google News Images post explains why the blending of Google News images into the Google Images doesn't quite answer that issue.
Posted by Danny Sullivan at 12:27 PM | Permalink
Red Herring reports that Perfect 10, an adult magazine and Web Site has filed a lawsuit in a Los Angeles Court against Google. The suit alleges that the, "search engine giant provided Internet users with at least 800,000 unauthorized links to images of Perfect 10s nude models, stealing membership fees and advertising revenue from the Los Angeles publisher. The lawsuit is one of the first of its kind against Google."
According to the suit, because Google profits from the misdeeds of others on the web, it is legally and financially responsible for the alleged violations.
Update: John Palfrey from the Berkman Center at Harvard Law offers an anlaysis. He has also made available a redacted version of the complaint (removing some "graphic" images).
MediaPost also provides more coverage: Adult Publisher Sues Google For Copyright Infringement
Posted by Gary Price at 8:31 PM | Permalink
A post at AnandTech raises concerns that Google Images fails to find pictures of US soldiers involved in the Abu Ghraib prison torture scandal. A conspiracy? In reality, just a failure of freshness on Google's part.
After the post was discussed at Slashdot, Google cofounder Sergey Brin sent word via Slashdot that Google was embarrassed to say its image database just hadn't been updated recently.
John Battelle summarizes more in his Google Image Search: Updated Only Twice a Year? post. While accepting Google's explanation, he wondered how the images that were there before could go missing. Others at Slashdot wondered as well. So did I.
Sure, information comes and goes on the web all the time. But for there to be no pictures at all for someone like Lynndie England, as this Google Image search shows? That seems odd, especially when the same search at Yahoo Image Search has plenty of examples (before forewarned, the link brings up several graphic and shocking pictures).
OK, Yahoo Image Search just recently got updated, as we covered two weeks ago: Yahoo Announces Size Increase to Image Database. But even if Google isn't as fresh, as it now readily admits, that still doesn't explain how the pictures that were once in there are now gone.
Answer? Google News. Google tells me that Google News has its own image search database that flows into Google Images. So when you search on Google Images, you're searching both the Google Images picture database (the stale one, currently about six months out of date) and the Google News image database.
Google won't say how long images stay in that Google News image database. I'd guess something like one to two weeks. That helps explain why a Google Images search for Lynndie England or Abu Ghraib might have brought up different results two weeks ago.
In October, Lynndie England gave birth. News stories about that may have brought pictures of her into the Google News database -- and thus into Google Images. Similarly, sidebar stories on the Abu Ghraid tortures or any other stories like this that may have hit Google news would have done the same. But as that news became old, those pictures fell out of Google News -- and out of Google Images.
The Google Images staleness prompted Yahoo's Jeremy Zawodny to suggest Google might want to outsource image search results from Yahoo. That's unlikely, but Gary Price does give you some fresh image search alternatives here.
Posted by Danny Sullivan at 2:57 PM | Permalink | Comments (0)
PixMisGoogled Sourc: Ongoing
Tim Bray provides an example of how Google Images "ignores" the alternatve text that's often associated with image files.
Posted by Gary Price at 3:05 PM | Permalink | Comments (0)
A post titled, "Amazon pushing porn?" dings a9 about their image search tool.
>From the post, "A search for an innocuous term such as "Frontpage SEO" will call up images of nude women on the first results page, alongside Web search results for Microsoft FrontPage software and links for data on search engine optimization (SEO)...People can filter 'mature' pictures from A9 or close the column altogether. But even with the image-filter setting on "strict," pornographic images may seep through."
Of course no image-filter is perfect but the issue might be more of a Google issue than an a9 problem. a9 uses Google's image database and their SafeSearch filter (it's mentioned on the a9 preferences page).
Btw, when I ran the "Frontpage SEO" search at Google.com today with "strict" filtering turned on and received the same/similar results that News.com found at a9.
Posted by Gary Price at 12:52 PM | Permalink | Comments (0)