Archive: June 22, 2003

<<< June 21, 2003

Home

June 23, 2003 >>>


Human Pattern Recognition

Sunday,  06/22/03  10:26 AM

Among all the remarkable things that we humans can do, effortlessly, that takes huge amounts of effort to do any other way; our image pattern recognition is perhaps the most awesome.


Original Image


Poor Compression


Difference


Good Compression


Difference

I was just struck by this in working with various compression schemes for images.  The best way for me to evaluate a compression routine is to load an image in Photoshop, then load an image which has been compressed and decompressed with the routine in question, and create a new image which is the "difference" between them.  I can then visually tell right away how well the compression routine worked.  This is much better and much more informative than writing a program to perform a detailed comparison between the two images.

Not only does this tell me quickly how well the routine worked - the degree of artifacts introduced by the compression - but it tells me what kind of artifacts are introduced.  Does the routine introduce false edges?  Does it attenuate areas of high contrast?  Does it smear detail?  Does it affect all color channels equally?  Does it have problems with very light areas, or very dark?  The only reason I can discern these things so easily is that my personal pattern recognizer is so good.  Human vision is amazing.  It can pick up edges easily, with no effort required.  It can detect subtle differences in contrast, across a very high range of intensity.  It can compare two regions of high-frequency detail and tell them apart.

If you look at the images to the right, you can easily see the "poor compression" is inferior to the "good compression".  Your pattern recognition is that good.  You might not realize why; but looking at the difference images, you have quantitative confirmation, as well as information about the types of artifacts (note all the little horizontal lines!)  Writing a program to compare the images and find these artifacts would be non-trivial.

As part of Aperio's approach to Automating Pathology, we've created a lot of amazing tools to perform pattern recognition on very large images.  These tools are much faster and much more quantitative than anything a human can do.  But their qualitative abilities are vastly inferior.  In fact, a big part of the technique we use is to automate the laborious pre-processing of images to optimize the use of human pattern recognition, just as "differencing" with Photoshop does.

 

Sunday,  06/22/03  11:46 PM

Tomorrow is the big day; Steve Jobs introduces Panther (the new version of Mac OSX) and the G5 PowerMacs.  I can't wait.

There were some Panther screen shots posted on the web, but they're gone now; Apple's legal team has intimidated everyone into taking them down.  It looked really nice, though...

And meanwhile, Intel is announcing a 3.2GHz P4 and will "discuss" a new version for dual-processor servers.  This is just a coincidence, right?  {Commentary: Intel's current top speed is 3.1GHz, so jumping to 3.2 hardly merits a press conference.  And discussing a future version of something is different from releasing it, or even from announcing it.  This is pure piffle.}

Bruce Sterling in Wired: New World Disorder.  "There are four ways to solve planet-wide problems.  None of them work."

  1. Global multilateral organizations.  The U.N., the World Bank, the International Monetary Fund, the World Trade Organization.  "They might look big and scary to street protesters, but once you peek behind the velvet curtains, it's dead obvious that they're stretched thin, put-upon, weak, fractious, crooked, and low in morale."
  2. International treaties and conventions.  "These vast, clotted webs of apparent consensus are too many, too messy, and too meager to manage a teeming, boisterous world."
  3. The coalition of the willing.  The Group of Seven Industrialized Nations, the Group of Eight, the Group of 20, and the weirdly named Group of 77 Plus ChinaBork.  "Coalitions of the willing are barely coalitions, they're rarely willing, and they're never broad enough."
  4. Glamorous international powwows.  The Rio Summit, Rio Plus Five, Rio Plus Ten, the Cairo summit on population, the Durban racism summit, the Copenhagen Social Summit, and, lately, nongovernmental countersummits like the World Social Forum.  "These massive blabfests are ritualized and wooden. They make proper noises, but they have no teeth, no budget, and no follow-through."

Unfortunately he is dead right.  The only thing that actually seems to move the ball forward is unilateral action by a benevolent power like the U.S.

All aboard!   A QTVR scene from the Hogwarts Express - a special steam train to celebrate the release of the fifth Harry Potter book, filled with 800 fans dressed for the occasions.  Wonder what Dumbledore thinks of this :)

Weird Al Yankovic interviews Chris Pirillo (links to video).  Scenes from a parallel universe...

Scoble compares Microsoft to Baskin Robbins.  (As in, who will be the next Cold Stone Creamery?)  Well, there's more to it, of course; please click through and read it.  Related - Mike Sax reports on Microsoft's Big, Big Problem.  "Developers are the ultimate king-makers in the platform wars."  If you don't believe this, look at what's happened with game consoles; the X-box is badly behind the PS2 solely because Sony made more and better deals with developers than Microsoft.

Billboard reports Top Artists Balking At A La Carte Downloads.  Interesting...  I guess these would be the artists who 1) don't get it, and 2) like to pad their albums with junk.  Imagine buying books this way - you can only buy this book as part of a set of ten, even if you don't care about the other nine.  Yeah, right.

The Wine Spectator reports French Wine Sales Still Dropping in United States.  Down 27% in the last four weeks.  Excellent.  What do you call this?  A good start.  [ via Instapundit ]

Evan Harris has published an interesting technique for spam blocking he calls "greylisting".  The gist of the technique is that the first time you receive email from someone, you fail to accept the connection.  If the sender is legitimate, they'll retry, and you'll receive the mail successfully.  Thereafter they will be "known" and you'll receive email from them straightaway.  If the sender is not legitimate chances are high they will not retry the connection, and thereafter you block all email from that source.  This technique takes advantage of the fact that most spammers use software specialized for sending spam which doesn't obey the "normal" rules of SMTP, which specify that you retry if a message cannot be delivered.  Fascinating...

 
 

Return to the archive.