IT-LIST Digest 36Topics covered in this issue include:  1) RE: IT counting bug?	by Christian Broesamle   2) Re: IT counting bug?	by "Don Wilcox"   3) matching moles in pictures taken 1 yr apart	by Greg Blair   4) Stacks	by Piero Bianco   5) Re: Stacks	by Fran\gois Richard   6) Stacks -> avi?	by Len Cleary   7) Re: matching moles in pictures taken 1 yr apart	by Robert Smith   8) TWIAN acquire	by James Zheng   9) Re: Stacks	by "S. Brent Dove"  10) Re: Beginner question - densitometry	by "S. Brent Dove"  11) Re: Beginner question - densitometry	by dcm@image.bio.bnl.gov (Denise Monteleone)----------------------------------------------------------------------Date: Sun, 11 Aug 1996 15:19:37 -0700From: Christian Broesamle To: "it-list@sparky.uthscsa.edu" Subject: RE: IT counting bug?Message-ID: <320E5C79.824@hifo.unizh.ch>Yesterday I wrote:>I experienced the following problem today: When i wanted to manually>count objects in an image with the Count/tag tool I didn?t get over 7>objects. The tool just didn`t continue counting and tagging anymore when>I clicked the mouse button. All other functions remained normal and I>tried it a couple of times with the same result (7). After some>frustration I exited IT and started it again and guess what? I could>count up to 9! so I had some hope that with the next starting up I would>maybe be able to count up to 11! But alas, it was only 8. This was when>I gave up and counted the stuff with the finger on the screen (using my>fatty fingerprints as tags :-)). Has anybody experienced a similar>problem and can help? I'm running IT under Windows NT3.51 on a 166>Pentium 16 MB RAM.Today I am one step further.I found out that the count/tag tool only works in a area of ca 420 x 340pixels in the upper left corner of my image (1280x1024). Because Istarted counting there this is why I never got over my 7-9 objects. Am Idealing with some predefined area of interest? If so, how can I countthe whole image? Or is it a bug?Greetings,Christian Broesamle-- ***************************Christian BroesamleBrain Research InstituteUniversity of ZurichSwitzerlandbroesam@hifo.unizh.ch***************************------------------------------Date: Sun, 11 Aug 1996 06:36:10 -0700From: "Don Wilcox" To: Subject: Re: IT counting bug?Message-ID: <199608111342.GAA27339@dogbert.xroads.com>----------> From: Christian Broesamle > To: wilcox@xroads.com> Subject: RE: IT counting bug?> Date: Saturday, August 10, 1996 6:38 PM> > Yesterday I wrote:> > >I experienced the following problem today: When i wanted to manually> >count objects in an image with the Count/tag tool I didn?t get over 7> >objects. The tool just didn`t continue counting and tagging anymore when> >I clicked the mouse button. All other functions remained normal and I> >tried it a couple of times with the same result (7). After some> >frustration I exited IT and started it again and guess what? I could> >count up to 9! so I had some hope that with the next starting up I would> >maybe be able to count up to 11! But alas, it was only 8. This was when> >I gave up and counted the stuff with the finger on the screen (using my> >fatty fingerprints as tags :-)). Has anybody experienced a similar> >problem and can help? I'm running IT under Windows NT3.51 on a 166> >Pentium 16 MB RAM.> > Today I am one step further.> > I found out that the count/tag tool only works in a area of ca 420 x 340> pixels in the upper left corner of my image (1280x1024). Because I> started counting there this is why I never got over my 7-9 objects. Am I> dealing with some predefined area of interest? If so, how can I count> the whole image? Or is it a bug?> > Greetings,> Christian Broesamle> -- > ***************************> Christian Broesamle> Brain Research Institute> University of Zurich> Switzerland> broesam@hifo.unizh.ch> ***************************This one we know of.  You are probably viewing your images at a reduced zoomfactor.  This is a known bug in 1.23, and has been fixed in 1.25, which istenatively scheduled for Aug 24.Don------------------------------Date: 	Sun, 11 Aug 1996 07:43:24 -0400From: Greg Blair To: it-list@sparky.uthscsa.eduSubject: matching moles in pictures taken 1 yr apartMessage-ID: <320DC75C.5A8E@dgp.toronto.edu>As part of a dermatology study, we have to match moles ina pair of patient pictures taken 1 year apart.We have to:1. identify moles that match up between the two pictures2. identify the moles that have gone away    (in image 1 but not in image 2)3. identify new moles   (in image 2, but not in image 1)As moles have a sharp transition from flesh tone to mole, they areeasy to classify.  Insect bites and acne have a much more gradualtransition and are easily rejected.The problems are:1. patients change weight over the year.    (extreme case: consider a 9 mo pregnant female)2. patients while standing in the same position, can have   different postures.  Arms and legs can be in different   positions and the trunk can have a different left-right    tilt.Neural network pattern matching has been suggested.  The problem is training the neural network on only 1 image before asking it to classify the second image.Does anyone know of any good matching algorithms?.. Greg Blairsiggraph@dgp.toronto.edu------------------------------Date: Mon, 12 Aug 1996 08:10:12 -0700 (PDT)From: Piero Bianco To: ImageTool Subject: StacksMessage-ID:         I am trying to create stacks in IT.  I create these first in NIHImage, then import/open the tif files into a new stack in Imagetool.  Thefirst time I did this it worked great.  SInce then Imagetool crashes assoon as I try to creat a new stack.  Am I doing something wrong, is therea problem with Image tool or I have I developed a conflict somewhere?  (Ido not believe it is a conflict as no new software has been installed andno system changes have been made).Thank you for your assistance.Piero Bianco------------------------------Date: Mon, 12 Aug 1996 13:56:38 -0400From: Fran\gois Richard To: it-list@sparky.uthscsa.eduSubject: Re: StacksMessage-ID: <320F7055.101C@geol.uottawa.ca>Piero Bianco wrote:> >         I am trying to create stacks in IT.  I create these first in NIH> Image, then import/open the tif files into a new stack in Imagetool.  The> first time I did this it worked great.  SInce then Imagetool crashes as> soon as I try to creat a new stack.  Am I doing something wrong, is there> a problem with Image tool or I have I developed a conflict somewhere?  (I> do not believe it is a conflict as no new software has been installed and> no system changes have been made).> > Thank you for your assistance.> > Piero BiancoI experienced this problem myself, and was told by Don Wilcox (IT God)that this bug had been identified and fixed in version 1.25, to be(tentatively) released Aug. 24.Cheers,-- Fran¨ois A. Richard            Email: richard@geol.uottawa.caUniv. of Ottawa (Geology)      Tel:   (613) 562-5800, ext. 6850Ottawa (CANADA) K1N 6N5        Fax:   (613) 562-5192=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= http://www.geol.uottawa.ca/geo/lists/grads/richard/richard.htm------------------------------Date: Mon, 12 Aug 1996 16:51:52 -0500From: Len Cleary To: it-list@sparky.uthscsa.eduSubject: Stacks -> avi?Message-ID: <9608122138.AA18055@nba19>Hello,Can anyone suggest a way to convert an IT stack to an avi movie? I don'tthink our machines have enough horsepower (100 MHz pentiums, 16 MB RAM) touse Imagetool as a Netscape viewer.Cheers,Len Cleary, Ph.D.Dept. of Neurobiology & AnatomyUniv. Texas Houston Medical School------------------------------Date: Tue, 13 Aug 1996 12:06:48 -0400From: Robert Smith To: it-list@sparky.uthscsa.eduCc: siggraph@dgp.toronto.eduSubject: Re: matching moles in pictures taken 1 yr apartMessage-ID: <1.5.4.32.19960813160648.006828f4@curtech.com>At 10:38 AM 8/11/96 +0600, you wrote:>We have to:>1. identify moles that match up between the two pictures>2. identify the moles that have gone away >   (in image 1 but not in image 2)>3. identify new moles>   (in image 2, but not in image 1)>The problems are:>1. patients change weight over the year. >   (extreme case: consider a 9 mo pregnant female)>2. patients while standing in the same position, can have>   different postures.  Arms and legs can be in different>   positions and the trunk can have a different left-right >   tilt.>Greg  Like most interesting problems, I'm not aware that this one has a "canned"solution.  1. How many of these do you have to do? A fully-automated system will bepretty expensive; not worth it, most likely.  COnsider a half-way solutionlike a program which puts the images side-by-side, and the operator pairs upcorresponding moles with the mouse.2. There is a lot of morphing software around, which will allow you tostretch an image in all sorts of ways. You could use this to alter one imageuntil various reference points (navel, tip-of-nose, etc) line up, at whichpoint most everything else will line up.  This would take care of changes inweight, and even modest alterations in posture (if you lined up the hips,and then the navel was off-center because one person was standing slightlysideways, moving the displaced navel would stretch one side of the torso andshrink the other.) An advantage is that morphing is very popular among"hackers" and you can get a reasonable quality of software cheap.  On theother hand, I doubt you can totally automate the process.3. I don't think it would be too hard to devise a program which -- given alist of apparent moles and their coordinates, and given a singlerepresentative pairing as a refer3ence -- would then move down the listpairing moles on the basis of size and a reasonable guess as to where themole would probably be.  It's hard to be sure how well such a simpleprocedure would work.  I should guess pretty well if you're not using peoplewith a veritable forest of moles.  I'm doubtful you'll ever automate thecase of the 9-mo pregnant woman. tho!Good luck, and don't hesitate to contact me again.Bob             .  Robert A. Smith, Ph.D.  _____    .    Vision Systems' Analyst |     |<.      Current Technology, Inc. |_____|   .    (603) 868-2270     ^       .  ras@curtech.com    / \   /   \------------------------------Date: Tue, 13 Aug 1996 12:51:33 -0400From: James Zheng To: "'it-list@sparky.uthscsa.edu'" Subject: TWIAN acquireMessage-ID: <01BB8916.2B4D6980@rwja.umdnj.edu.umdnj.edu>Hello,I understand that IT 1.25 is to be released soon (August 24?).  Can you =guys modify the program so that TWAIN Acquire will not AUTOMATICALLY =close after acquiring an image.  I am using TWAIN interface to control =an image processor (Hamamatsu Argus-20) and I would like to leave the =TWAIN acquire open until I finish all the work.  Thanks.As for writing customer plug-in, how do I deal with 16-bit images (i.e. =from cooled CCD).  ITSDK mentions that the acquire plug-in should return =a DIB, but I have no idea about 16-bit grayscale images.  I plan to =write a IT plug-in for Photometrics PXL series of CCDs. =20James Zheng, Ph.D.UMDNJ-RWJMSDept. of Neurosci. & Cell Biol.web: http://www2.umdnj.edu/~zhengjq------------------------------Date: Tue, 13 Aug 1996 17:00:38 -0500 (CDT)From: "S. Brent Dove" To: IT List Subject: Re: StacksMessage-ID: <01I88GCKMD2C005LK3@uthscsa.edu>Recently on IT List>        I am trying to create stacks in IT.  I create these first in NIH>Image, then import/open the tif files into a new stack in Imagetool.  The>first time I did this it worked great.  SInce then Imagetool crashes as>soon as I try to creat a new stack.  Am I doing something wrong, is there>a problem with Image tool or I have I developed a conflict somewhere?  (I>do not believe it is a conflict as no new software has been installed and>no system changes have been made).Are you using version 1.23 of ImageTool.  We did discover a bug in the create new stack in which a recursive loop occured when creating new stacks (version 1.23).  We have since fixed this bug in version 1.25 which should be out by August 24. Thanks for letting us know about this.S. Brent Dove                            Voice: (210) 567-3333Diagnostic Sciences                      Fax:   (210) 567-3334University of Texas                      Email: dove@uthscsa.eduHealth Science Center                    Web:   ddsdx.uthscsa.eduSan Antonio, TX    USA                   ftp:   maxrad6.uthscsa.edu------------------------------Date: Tue, 13 Aug 1996 17:00:43 -0500 (CDT)From: "S. Brent Dove" To: IT List Subject: Re: Beginner question - densitometryMessage-ID: <01I88GCOAG2O005LK3@uthscsa.edu>Recenlty on IT List>I just started using image tool. The "line profile" provides a useful>densitometry of a gel, but I would prefer using a 'thicker line' or>a rectangle, and average the pixels to obtain a more representative>densitometry.  How can I do that ?You speak, we listen.  We have added this feature in version 1.25.  You can specify the number of adjacent line profiles which will be averages to get better data for densitometry of electrophoresis gels.S. Brent Dove                            Voice: (210) 567-3333Diagnostic Sciences                      Fax:   (210) 567-3334University of Texas                      Email: dove@uthscsa.eduHealth Science Center                    Web:   ddsdx.uthscsa.eduSan Antonio, TX    USA                   ftp:   maxrad6.uthscsa.edu------------------------------Date: Wed, 14 Aug 1996 09:22:52 -0400From: dcm@image.bio.bnl.gov (Denise Monteleone)To: it-list@sparky.uthscsa.eduSubject: Re: Beginner question - densitometryMessage-ID: <199608141322.JAA11959@image.bio.bnl.gov.>Brent,	I am really glad that you are adding more features for electrophoresisgel analysis. We do alot of them and also sequencing gels. So thanks, and keepthoses new features coming.					Denise Monteleone					Brookhaven National Laboratory					------------------------------End of IT-LIST Digest 36