Re: fingerprint recognition

"Chaos" <rothko.fan@xxxxxxxxx> wrote this uninformed and insulting message :
this "ImageAnalyst" appears to have no real world experience, if you look at his posts, he always refers to using toolboxes. in one post, he had no clue why someone would recast an image to uint8. i think he's just someone fresh out of college matlab hired to post replies in the newsgroup. all he does is quote papers and toolboxes to get people to buy toolboxes. he solution is basically out of book.
go to the website and scroll down and look at
"Fingerprint enhancement"
all the routines are there and ready to go. his method will work IF you have the toolbox. these routines require NO additional routines and are well documented.
----------------------------------------------------------------------------------------------------------------
----------------------------------------------------------------------------------------------------------------
----------------------------------------------------------------------------------------------------------------
Chaos:
Your comment is so absurd and presumptuous that I had to decide whether to deign to reply, since a reply might indicate that I even care what you think. However your public attempt to besmirch my reputation compels me to respond lest others believe the seeds of doubt you have so ungraciously sown. The fact is, I have lengthy and extensive experience in optics, image processing, and image analysis. Maybe your reply is just a snarky way to extract that out of me, but since you forced the opportunity....

I have a Masters and a Ph.D. in optics from the top school in the field. My doctorate is in medical image processing. I have 10 years of academic experience in image analysis, 2 years of military imaging experience, followed by 21 years of industrial imaging experience. I currently work in the imaging division of one of the world&#8217;s largest global corporations, one that virtually everyone on the planet has heard of. I have taken courses and taught courses and seminars in image processing and color science, and continue to do so. I go to image analysis symposia, publish papers, present posters and talks, have industry connections, and contribute to several imaging-related newsgroups. I give seminars to other corporations, universities, and internal to our company. I&#8217;ve given seminars at government-affiliated labs like Lincoln Labs and JPL. I've written updates to ASTM
standards. I've refereed papers. I'm called upon as an imaging expert in litigation (one right now as a matter of fact). I'm frequently called upon to write and describe the imaging parts of patent applications. I'm an "inventor" on patents. One of my patented applications has been featured on national news shows, local news shows, talk shows, news magazine shows, and in several print magazines. In fact you might even have used it yourself.

I have worked on literally hundreds upon hundreds, if not thousands, of image analysis projects. That is no exaggeration, repeat no exaggeration. At any given time I am working on 10-20 image analysis projects. I design optics, imaging rigs, algorithms, and design and write end-user software run by hundreds of users. We spend hundreds of thousands of dollars per year on systems I design for internal use. I deploy applications to numerous research centers on 4 continents for use by fellow scientists in my company. I travel the world to teach, install imaging systems, and train users how to use them. These are all real world situations complete with blur, noise, color correction, colorimetric calibration, spatial calibration, background correction, intensity calibration, etc. I plan monthly imaging seminars and two to three larger symposia per year. I'm in charge of an imaging web
site serving 700 scientists in our company who use imaging. I work on an incredibly vast variety of projects - it's not like all I do is one thing such as astronomy, semiconductor inspection, medical imaging, or microscopy. Few companies in the world (you could probably name them) offer opportunities to work on as a wide variety of applications as I do. I work with dozens of different imaging modalities (in spectral ranges from x-ray to thermal IR) in collaboration with experts in those areas. My current list of projects includes images captured by MRI, CT, a microscope, a line scan camera, a profilometer, a monochrome video camera, X-ray diffraction instrumentation, 2D digital X ray transmission images, an ultraviolet camera, and color video cameras. And those are just my ACTIVE projects this month. Apparently the other scientists in this corporation think I'm qualified enough to
be a top-rated senior level scientist responsible for the imaging-based performance testing of several products which each earn us billions (yes, billions) of dollars per year. Every word of what I just said is the absolute truth. It would be pretty damn amazing to accomplish all that just a few years out of college.

Now my 30+ years experience with an incredibly wide diversity of imaging modalities and algorithms may not make me the most experienced imaging guy out there, but I know I'm knowledgeable and experienced enough to answer some of the imaging questions posted here. If someone posts a question that I think is simple, fun, or interesting, then I may post some sample code to help them, preferably on one of their images but otherwise on a standard MATLAB demo image (sorry - but some of them come from the image processing toolbox). Oh, occasionally I might do some good natured teasing of a poster for not posting their image, or texting their message out on a cell phone, but that's the spirit of this newsgroup. But if a user posts some half-baked question or statement such as "I need help finding faces in videos" then I'm not going to write a complete turnkey application for them. Neither
would you. The best we can do is to point them to some other references such as web sites (for example the renowned USC bibliography web site, or other researcher's web page) where they can find more information, publications or algorithms, or perhaps even code. Then they can take it from there. Maybe they will come back with a more targeted question. You do the same thing - many of your responses simply tell them where to look in the MATLAB help file or point to a web site (such as the one on fingerprint analysis). We're all too busy to hold their hand for very long (although Dave Robinson will often give it a good run). So your inaccurate and hypocritical comment about me referring them to other places is like water off a duck's back. It won't change my policy a bit.

People in this newsgroup know the quality of the information I provide. Several people at the MathWorks have met me in person, know me, and respect me but they do not pay me. I recommend their products because they are good products. I'm confident and secure in the reputation I've earned here. People here know who knows what, and who provides quality answers. I've only been using MATLAB for several years and know only a fraction of what respected gurus like Walter, John, Steve, us, Rune, Bruno, Matt, TideMan, and many others know about MATLAB in general, but I DO KNOW IMAGING. It's my life's work and hobby. I&#8217;m obsessed with it. Ten to twelve hours a day seven days a week. I spend several hours a week reading and learning new algorithms. I eat, live, and breathe imaging. It&#8217;s in my blood. I can't get it out of my head. I even dream about it. It IS me. Imaging,
optics, and image analysis are all I&#8217;ve ever done in my 30+ year professional career. You earn and deserve the reputation you get, and I'm happy with mine. I'm sure Dave Robinson (one of the imaging gurus here) is also happy with his reputation as he has shown himself to be very generous and helpful. Likewise for Walter, John, Oliver, and others who regularly answer imaging-related questions here. But I'd like to keep my reputation positive and your comments could damage that unless people were familiar with both of us. Now, I don't know what experience you have and I won't rudely and incorrectly speculate on that as you have done with me. You may have more experience than me - I don't know - but that would be fine. In fact I'm sure that you're better than me at your particular specialty. So I'm not going to impugn your experience any more than I would impugn John's or
Walter's (but you're welcome to - oh yes, that would be SO fun to see). I&#8217;m not going to assert that you are some paid inexperienced shill fresh out of college like you did to me. I have no basis for saying that. You are earning your own reputation here based on your posts, and you're welcome to it. Most have been helpful but others (one case in point above) have, in my opinion, been snide, sarcastic, or overly sensitive remarks that I've held my tongue over (although some haven't - http://tinyurl.com/dxnc8l). Before you compelled me to speak out, I let your posts speak for themselves. If people want to take you up on your good suggestion of looking up my past contributions and compare them to yours, then that would be fine. Yes, I would be QUITE fine with that. Many, many people thank me for my help (e.g. http://tinyurl.com/dxoq95, http://tinyurl.com/dd8cwv, etc., etc.).

Your comment about uint8 is such utter nonsense I won't even dignify it with a response. So...moving on,...I didn't say that the poster couldn't go to the web site you recommended and try out that fingerprint toolbox/suite. I merely gave one way of solving it. I don't always have the best solution - Walter, John, and others frequently give a better answer than I gave and I never take offense at that. I don't get mad at them or the poster (http://tinyurl.com/dj68n8) for not going along with my suggestion. In fact I learn a lot from looking at their informative posts. I love their contributions and look forward to them. Now, we know you're anti-toolbox, and proudly claim about this fingerprint toolbox "all the routines are there and ready to go. his method will work IF you have the toolbox. these routines require NO additional routines and are well documented. " ---- Oh
really?????? ---- Well . . . I looked up the code and it DOES REQUIRE THE IMAGE PROCESSING TOOLBOX. For example, there are calls to Image Processing Toolbox functions fspecial, imrotate, and ordfilt2. Sure you could replace those with your own hand-written versions but it's not "ready to go" without toolboxes. Granted, it is well documented and a very nice collection of routines. Thanks for pointing it out (apparently you're allowed to point people to web sites but I'm not).

And what's your problem with toolboxes? Several of your posts negatively mention "expensive toolboxes." Your response above draws the absurd conclusion that anyone who always uses toolboxes has &#8220;no real world experience&#8221; and is insulting to those of us who use MathWorks toolboxes. The MathWorks provides them so why not use them? Frankly it's a waste of my time to rewrite basic things such as labeling, regionprops, morphological operations, etc. If you want to, go ahead, but I have better things to do. Not everyone is so cash-strapped that they need to write everything themselves from scratch. I'll take a small fraction of our multi-billion dollar R&D budget and save myself some time. Steve Eddins&#8217; group at The MathWorks makes a fine toolbox that enables me to be more productive on higher level concepts. You rail against toolboxes, yet you then suggest to a
poster to go use one on fingerprints (which even requires a MathWorks toolbox). In fact, I think it's also hypocritical for you to even use MATLAB. That's just another toolbox of a sort. If you're so against packaged code, why don't you just write all your own code in C or Java instead of using MATLAB at all? Why use MATLAB to multiply matrices when you can do it yourself in 5 lines of C? The Image Processing Toolbox must be one of the most used toolboxes in MATLAB. LOTS of people have it! Just look how many posts there are on images. I doubt many people here agree with your anti-toolbox philosophy. Often I will give non-IPT options, such as hist instead of imhist. There is nothing wrong with asking if a person has the Image Processing Toolbox or offering up code that uses it, and I'll continue to do it.

The next time you think about writing something like that about me or someone else, stop and think. It might really say more about YOU than about the person you&#8217;re trying to slam. Unfounded speculation can backfire.
Signed,
ImageAnalyst
P.S. I believe that, technically, you&#8217;re probably an experienced, bright, talented, and clever engineer and I look forward to seeing some informative and respectful contributions from you in the future. I never harbor hard feelings or burn bridges because learning and excelling at my projects are too important &#8211; perhaps you will help me with a question some day.
.

Relevant Pages

  • Re: Syracuse Convention warning
    ... The website doesn't have it listed, but I'm going to be teaching a ... I am going to need 5 volunteers for one of the new routines. ... Check out our much more informative site ...
    (rec.juggling)
  • Re: Re-Attach Tables
    ... You'll probably find multiple references to routines that do this if ... website. ... tables if for some reason they were removed from the access ...
    (microsoft.public.access.tablesdbdesign)
  • Looking for template code for Lapack++
    ... website ... I'm interested in the routines which are ... for complex valued matrix is called. ... using namespace std; ...
    (sci.math.num-analysis)
  • Scrolling the ZX81 screen in assembly
    ... I've added routines to scroll the ZX81 screen to my website: ... Have fun. ...
    (comp.sys.sinclair)
  • RE: get ip address
    ... They have other similar routines, so you may wish to search a little further ... through their website. ... Daniel Pineault ... "iccsi" wrote: ...
    (microsoft.public.access.formscoding)