ImageJ's Thermal LUT mode for generating a heat map largely rids the TS face of its alleged "banding effect", prompting reinterpretation of the facial cut-offs.

Note added October 7, 2016: Dear reader. Please be aware that my 5 year Shroud investigation has come on considerably since penning this post over 2 years ago. Indeed, one could say it's reached a conclusion - a very definite one - summed up in this recent photograph:

Centre: plastic Galaxy Warrior, approx 1/12 human scale. Left: thermal  imprint of the same onto linen, using oil/flour as imprinting medium, followed by oven-roasting of the imprinted linen. Right: the same thermal imprint after tone-inversion and 3D-rendering in ImageJ. Note the amazingly close correspondence with the original figurine, except obviouslyfor colour.

Yes, I believe the Shroud to be of medieval provenance, consistent with the radiocarbon dating (1260-1390). Moreover, I believe it to be a thermal imprint of a real adult man, obtained using the technology decribed above. A final wash of the thermal imprint with  soap and water would have removed all traces of the imprinting medium. That would leave just a faint attenuated 'enigmatic' image, a fair description one might think of the Turin Shroud. Purpose?  The latter was intended to represent what a 1300 year old body imprint (sweat, blood) would or might look like some 1300 or so years later.

See my recent Shroud posting for more details.


Start of the original 'Thermal LUT' posting:

How many times have we been solemnly assured that the sharp cut off at opposite sides of the TS image is "just a banding effect" due to yarn-to-yarn variation?  How probable is that given the high degree of bilateral symmetry, or the break in linearity at one of the cheekbones (your right)? See my previous attempts to get my head round the existing and largely unsatisfactory literature.

Don't ask (yup, there are so many things one is not supposed to ask, unless one wishes to invite ridicule - or  derogatory remarks that question one's scientific credentials).

Here's an image obtained using my new toy (hat tip to OK on t'other site) that should by rights kick that banding idea into the long grass.

Click to enlarge

(Technical note: the first graphic above was a Shroud Scope image to which I had given my customary contrast-enhancement in MS Office Picture Manager. For the Thermal image above, the input was Shroud Scope's "as is" image: the software seems perfectly happy to work on an entry-level image that is available to everyone, including those who don't wish to be bothered with photo-editing programs, and it's not a bad thing anyway to dispense with any initial image-enhancement that might risk creating artefacts, or lay one open to the charge of being cavalier in that respect).


It's been obtained by using a setting in ImageJ known as Thermal LUT ( LookUp Table) mode, available as an option on the 3D image-rendering option. It not only reads 2D image intensity, plotting that on a third dimension (z axis). It also colour-codes the gradation of image intensity - see chart on right - to create an admittedly somewhat fanciful  'heat map' ranging from blue (lowest relief) through shades of green and yellow to red (highest relief).

Straightaway one can see that the sharp cut-off at the sides of the face, aka excessive and unrealistic 'linearity', disappears almost completely.   With it must surely disappear any idea that the cut-offs had anything to do with banding. Instead, the cut-off must surely be telling us something about the nature of the 3D (or semi-3D) subject from which the (negative) imprint was taken.

(Late addition, Aug 31 : some have got the wrong end of the stick, and think I'm saying that banding is non-phenomenon. I've said nothing of the sort. Banding is a reality, though whether it's the result of yarn variation due to batch-to-batch variation in bleaching technique is another matter. See the 'new paradigm from Hugh Farey which attributes banding to interactions between the herring bone weave and mechanical bunching caused by rolling and folding etc. Nope. I'm not rejecting banding at all. I'm saying that the severe cut offs at the sides of each cheek, producing those pale largely image-free vertical bands between cheek and hair on BOTH sides, i.e. symmetrically, is nothing to do with  banding, and that other explanations must be sought. One does not need to agree with my explanation for the cut-off effect (properties of an imprinting template, not linen) to reject banding as the reason for the cut-offs.)

My own fast-crystallizing thoughts are the same as those expressed by the admirable Professor Luigi Garlaschelli, he of the imaginative and scientifically-sophisticated  reconstruction of possible medieval "forgery" technology, namely that a  bas-relief  template was used for the face, ie. a casting with shallow relief, but other possibilities are not excluded.

Here's an example of a modern-day bas relief made in bronze.  

Yup, that's heat-resistant bronze. How I wish I had something similar. It would be inside my oven before you could say "negative thermal imprint" or even "scorch".

Are there other features in that 'heat map' that fit with a bas relief template? Yes, methinks there's at least one. It's the intense imaging of the beard, and (especially) the moustache. In an thermal imprinting ("contact scorch") model, it's possible that a heavy imprinting of the chin might look like a beard. But a moustache also? Wouldn't that imply a row of front teeth that were set too far forward in the jaw bone?

One wonders how the "radiationists" would try to account for the prominence of facial hair in their fanciful models, those I tag as "theo-physics".  Wouldn't any hair be instantly singed off by any radiation sufficiently energetic to leave an imprint on linen?  Heck, how come those radiationists able to get away with describing themselves as "scientists" in their 'look-at-me' press releases, the same ones who bang on about the impossibility of a forger being able to create "microscopic complexity" at the fibre level, thus betraying their total ignorance of physical and chemical principles (reminder: atoms and molecules are self-ordering, almost as if they had a "mind of their own" in common parlance, which is needless to say a consequence of those quantized electron energy levels that spontaneously generates  'microscopical complexity'. See my previous analogy with the snow-making machine, where microscopic complexity arises simply from cooling water vapour or liquid aerosols, i.e. from abstracting, not adding kinetic and thermal  energy.

It should be clear by now that I have lost patience entirely with those radiationists. They are turning science and scientists into laughing stocks, especially when they preface their remarks to MSM journalists and bloggers with "speaking as a scientist". "Speaking as a totally gormless numpty" more like it.


I'll be looking at the rest of the TS image, frontal and dorsal, region by region with my new toy in the next few days. Maybe there will be more surprises and (hopefully) new insights in store. Thanks once again, OK.

Wednesday Aug 27

Here is the region of the crossed hands (a promising one in a diagnostic sense for distinguishing between contact-only v non-contact permitted models) shown at four levels of "heat", i.e. 3D amplification:

Click to enlarge

It's that first one that I find especially interesting, one in which the "heat" has been turned down to a very low setting, such as to show small areas of blue ("very cold") which might reasonably be interpreted as the basal background value of a non-imaged area. (Reminder: it would be unrealistic to assume that non-imaged areas are totally colour-free, since the TS background colour is a faint yellow, not a pure white.)

Click to enlarge

What we see there is a clear discrimination between sizeable areas of crossed hands/abdomen compared with surrounding areas, notably the thigh (viewer's right) and the areas immediately abutting onto the crossed hands and abdomen.

Can any conclusions be drawn?

One has to proceed cautiously, in view  of the background colour problem . But here's where the ability to ramp up the "heat" comes into its own, offering a superior analytical tool when compared to standard ImageJ 3D enhancements that lack the 'tunable' colour coding.

Let me explain. In the picture above, one of the thighs (reader's right) appears to be essentially non-imaged, which initially came as a surprise to me, and made one wary.Surely a thigh would be imaged, whether the imprinting was by contact-only or not?

However, it was imaged, as can be seen by turning up the heat, and its 3D shape becomes more obvious at the higher settings. So even a raised part of a subject's relief can be poorly imaged, relative to other raised parts (crossed hands). Is that because of an air gap between cloth and subject, which may have allowed weak imaging, or is it because there was no air gap, allowing direct contact between cloth and subject, but with some other factor operating to produce weak imaging, relative to those crossed hands and forearms?

More to come (after I've given those questions some more thought).

07:50   Well, I've thought. Frankly, I cannot see why the thigh is poorly imaged in any model that allows for non-contact. Either the thigh would be very close to the subject, or would be in physical contact, so the cloth-distance measurement is small  - a matter of mm at most- or zero, and if radiation were the only factor, then thigh imaging should have been stronger. That;s assuming that the radiationists are not suggesting that contact opens the door for non-radiation physics too. What would that be? Heat conduction? But that required a hot subject - like a heated statue. It is inconceivable that a corpse could ever be hot enough to imprint an image by heat conduction.

So where does that leave us? Let's forget about radiation and focus on heat conduction, requiring direct contact. But while direct contact in that model (the only realistic model in my view) is necessary for imprinting an image, it may not be sufficient to guarantee a strong imprint.

What are the other factors, apart from obligatory contact, that make for a stronger or weaker imprint? Model studies provide immediate answers. They are: 1. temperature of the template 2. contact time. 3. contact pressure  4. curvature of the template, resulting in different angles of presentation.

Let's home in some more on the fingers. They seem to show promise as an especially diagnostic area for distinguishing between alternative imaging mechanisms (radiation projection v direct imprinting) reinforced by some recent experiments I reported using imprints of my own hand. (It wasn't the entire finger that imprinted, only the bony central spine, giving the appearance of separated fingers when  in fact they weren't).

I started using the same 4 steps as used above for the crossed hands. Here they are in sequence:

.

.

And here for good measure is a 5th in the tuning-down sequence, with plenty of blue to show we are near to background levels of image intensity.:

Interpretation?  I'll be back when I've had time to digest.

Update: this scurrilous, indeed slanderous comment has just appeared on Troll Central:

anoxie

August 27, 2014 at 4:15 am

Reply

You’ve already answered Dan. This is just another obvious illustration of CB’s bad habit to distort/select/ignore facts.

 And what' prompted it you may ask? It's on a pirated version of this very posting, yet one instance in scores of my content being used within hours or even minutes of me posting to provide instant entertainment to the bored souls who assemble at that site to be given their daily fix of pro-authenticity reassurance.

Note the item I have ringed in yellow, top right hand corner. It's a link to the second of Thibault Heimburger's pdfs ( "potpourris de fantasie" as I call them). The first, also in the sidebar, claims that scorch imprints will always display excessive contrast compared to the image TS, to which my answer is/was: "Yes, they will if you choose an unsuitable template, i.e. a sunken relief instead of bas relief".  That claim, based on an cavalier piece of experimental design is now stated as fact in the wiki entry for Shroud of Turin under the sub-heading "Bas relief". (Yes, a wiki contributor too has got it wrong too).

As for that circled pdf, the second, words simply fail me, claiming as it does that one cannot scorch a few superficial fibres in a linen thread - that it's "all-or-nothing". That's not just wrong. It's laughably and grotesquely wrong, as I have pointed out, with and without a microscope, yet the pdf is steill there, being advertised, indeed highlighted in red. I've asked Thibault Heimburger to withdraw it - he hasn't. Yesterday I asked the site's owner to at least provide a facility for attaching comments to pdfs that are otherwise closed-off to criticism, and been met with a blank refusal. Meanwhile, that same owner continues to pirate my content on a daily basis.

What am I going to do? Answer: nothing. I've spelled out clearly what is happening, so as to let folk see things for what they are. The promoters of pro-authenticity have no ideas of their own - good or bad. All they can do is grub around for other people's ideas, and then sit in their little cave gnawing away at the meat and bone until all that's left is detritus. Far be it from me to deprive them and their provider of their sad, saprophytic* lifestyle.

* Saprophyte: any organism that lives on dead or decaying organic matter.

(Yup, the second part is an apt description of what happens to my postings  in that cave, after they have been messed about to make Porter's potted versions). 

Back to the pure clean air of science: at the risk of being accused of "selecting"  data (what could possibly have been selected - or rather de-selected- in this posting?) the images above dovetail neatly with those I produced last month from imprinting my hand:

From the archives: 

It seems so obvious now (why the Man on the Turin Shroud has bony fingers and no thumbs). The image is NOT a photograph. It's an 'impactograph'.

 

from Tuesday, July 15, 2014

The conclusion I arrived at then is reinforced with the new thermal graphics, inasmuch as the latter look "bony". Now why should that be, one may ask? What is it telling us about the imaging mechanism. Leaving aside those fanciful ideas that a  recently-deceased cadaver in a 1st century tomb turned briefly into a hospital X-ray machine (yes, really!) let's ask which fits - a radiation-allowing-air gaps model or a  conduction/contact only-no-air-gaps-allowed model.

Stretch out your hand, palm down, and place the edge of a ruler across the fingers. What the greatest gap between flesh and ruler? Answer - a few millimetres at most, so any conjectured radiation has only to cross a tiny air gap, where such gaps exist, everywhere on the hand. There is little potential there one might think for getting a clear impression of fingers, much less seeing prominent joints (knuckles) or gaps between fingers.

Now consider a contact-only model, There one expects all-or-nothing imprinting, so one DOES expect to see individual fingers, through failure to image between fingers where there is loss of contact. One might also expect the central bony part of fingers and the bony joints to imprint better than the more fleshy parts. Well, that is precisely what one is seeing in those thermal images, and precisely what one is able to model from one's own hands, coated with a safe and suitable paint ( I used a well-known brand of chocolate/hazelnut spread).

Selected data? The only data I select for my science is that which puts my readers into my garage and/or kitchen so they can picture with their own eyes what I am seeing and reporting. Why would I, with a lifetime in scientific research and education, be spending all this time if it were simply to deceive people? Who in their right mind would want to round off their life's work with a reputation for lying and cheating?

The only results I hold back are the ones that fail to make a point. I don't select points according to whether they seem to fit or not fit with a preconception. Scientific experimentation  starts with clearing one's mind of preconceptions.  Only a total moron would attempt to deceive himself and others. This blogger/retired scientist may be many things, some of them possibly unflattering, but he is most certainly not a moron. At least, my wife and grown-up children say I'm not, which is good enough for me. What that troll chooses to think is her business, but she ought to be more careful about what she says about people like myself who blog under their real names with reputations at stake, while hiding as she does behind her  French pseudonym. Why does Dan Porter allow that troll to trickle her poison onto his site? Does he think it spices up the site? If so, then he is equally culpable in allowing it  to be used as a platform for making character attacks. I'm not issuing threats - simply asking for a little human decency.

Update:  12:40. Comment addressed to the toxic anoxie:

David Goulet

August 27, 2014 at 7:27 am

Reply

Can you provide a Top Five list of CB’s Distorted/Ignored/Select Facts? If you are going to make a claim like this you need to back it up with ‘facts’ yourself. Otherwise you are guilty of one CB claim – that you are a troll ..........................

Thanks DavidG. It's good to see there are still folk like yourself, prepared once in a while to put their head above the parapet in defence of fair play.

13:45

OK, it's time now to go back to the face (Hugh Farey having made his usual perceptive analysis elsewhere, to which I may return later) and starting with maximum "heat", repeat the same process of draining away the colour-coded '3D-ness' to see what disappears first, and what remains at the end.

I'll be back later, hopefully with some succinct interpretation, or, there again, maybe not.

14:00 Toxic anoxie is back, responding to DavidG's challenge essentially to put up or shut up with a link to Thibault Heimburger's first pdf, the one with the crazy template that I mentioned earlier.

In fact I produced critiques of that pdf in short order - in three instalments - and when they were all ignored by Dan Porter (thus giving TH an easy ride) I did a post purely to flag up the existence of my 3 postings.

Here's a screen grab from 2012:

Here's a link to that paper:

Still Dan Porter ignored my critique, so I gave up and did other things. But the pdf in question remained in the sidebar over there, gradually acquiring hits over the months, gradually improving its Google ranking, so it's no surprise (1) that wiki has made use of it, despite its flawed experimental design and invalid conclusions and  now (2) toxic anoxie is attempting to deploy it as a proxy weapon against me, presumably having nothing of her own to offer.

15:30

I briefly toyed with the idea of posting this image to Troll Central, with the caption: "Spot the bas-relief template" in Thibault's first pdf:

That's bas relief as in this wiki definition:

A bas-relief ("low relief", French pronunciation: ​[baʁəljɛf], from the Italian basso rilievo) or low relief is a projecting image with a shallow overall depth, for example used on coins, on which all images are in low relief.

"As used on coins" note. The relief rises out of the base plane. it does not sink in from the base plane. It is also supposed to have smooth rounded contours, no ledges.

A proper bas relief penetrates into linen easily initially, since the highest planes account for a small farction of the contact area. Pressure builds as more and more of the planes enter the cloth. Thibault's template encounters high resistance as soon as it meets the clothe, and it will then be harder to imprint from the lower plane. Is it any wonder that his scorches display what he called "excessive contrast". It was an even greater shame that he went on to generalise, claiming that all scorch imprints would have excessive contrast. It's an even greater shane that the pdf is still on display on the Porter site, and that it's attempt to dismiss scorching  based on excessive contrast (or colour gradation) is now being relayed to the big wide world via the wiki entry on the TS.

But as regards posting the above to that highly problematical site (in more ways that one, given its tolerance for trolls AND promotion of dud pdfs, I quickly thought better of it. Why bother? It pirates enough of my content as it  is. Why go there simply to work a treadmill, endlessly correcti g all the same old errors and misconceptions. there are folks over there who need their shroud authenticity like others need Maltesers or TV soaps.

16:40:  I have a confession to make. I've been raising and lowering the "heat level" in the TS images without saying how it was done, and/or whether it was scientifically meaningful. What's more the way I did was discovered purely by twiddling controls in ImageJ without any thought as to the underlying matrix transormations (not that I recall much about those, except from private tutoring 'modern maths' some 30 years ago.

Draft only- may be changed.
So what control was used? It's actually one I rarely of ever use.It's the "max" and "min"  slides, bottom right of the ImageJ control panel, which I usually leave set on 100% and 0% respectively.

To produce the changes you see above, I varied the "min" value between about 0% and 50%, leaving the maximum unchanged.

How best can I demonstrate to my readers what changing the min value does? For now let's do it for a home-made image from a previous posting of mine that explored  ImageJ software, one with no 3D history, but a coding for 3D as used on old-style relief maps.

Here's the inputted  image:

Here's how it responded previously  in ImageJ without the new-deployment here of the thermal/colour-coding option.

The max and min controls were set, as I say, at 100% and 0% respectively, as can be confirmed by clicking on the image to enlarge, and reading off the two slide scale values lower right.

So let's enter the 2D image into "Thermal LUT" and see what we get, starting with  the 0%  min value:

OK, compare and contrast with the preceding image. They are comparable, thank goodness, with the bonus now of ImageJ's own colour coding of relief that overrides my own.

The next step is to ramp up the min value by degrees and see what happens in the model system, hopefully to better understand what was happening earlier in my 'suck-it-and-see' MO (this being  a quirky stream-of-semi-consciousness, age-impaired  real-time investigation site - as you may realize - as indeed it has been from Day 1, back at the tail-end of December 2011.

OK, that min value is now at 14%.

The change is subtle.(I'll try to articulate it  later when there's more progression, more trend data to work with.

We're up at 33% now. The effect we are seeing seems to be one of amplifying the highest relief, making it taller, but lopping off the highest relief. That's the point where I stopped with the TS image - when I saw those flat plateaux appearing. We'll now do one more step, if only to confirm the trend.

That's now at 76%.

Methinks it time to get back to the drawing board, where I will play around some more with Thermal LUT.  I'll switch to using my contour line model  probably returning to standard settings on max and min, given the complexity and subtlety of the response to changing the "min" value. (It doesn't  matter how dramatic the effect on the TS image if one cannot adequately explain what the software is doing).

Instead I'll try altering some of the  other controls. One is on a whole new learning curve here, but definitely worth exploring further to see if the software can unmask hidden detail and differences in the TS image.



Getting to grips with this thermal software may take some time, and I don't want to waste people's reading time with any more dead ends. I'll be back once I've made  sense of it. That may take some time, days, possibly weeks. Bye for now.

Update 21:40

David Goulet commented on Banding? Is it Real?.

in response to Dan:

Can anyone explain how the image* on Colin Berry’s blog can begin to convince us that banding is not really all that real. Maybe you can understand what Colin is saying. Something about “bilateral symmetry.” If anything, it helps to convince me that there really is banding there. You really need to see it in […]

Excuse my need for a general refresher here. The banding is important because, if it exists, it is problematic for the scorch theory because there is no way an artisan could have achieved this effect, therefore it must point to some other image formation process. Is this the gist of why banding matters?

Colin’s investigation points to another possible cause for the banding effect — it not being banding at all. Anoxie is countering that the banding is obviously there. Hugh is not ruling out banding, but is not as convinced as Anoxie as to it’s “obvious” existence.

Fair summation?

This blogger's response: 

My position, for what it's worth, based as much on arts and crafts as it is on science, is that the TS face looks quite unlike a real one - it's too rectangular, too gaunt, too "gothic".

I think the pro-authenticity folk know that, and have been desperate to provide a comforting albeit distracting narrative. It's based on the phenomenon of banding - bleaching differences in the yarn used for weaving.

Nobody doubts the existence of banding, but does it provide an explanation for the rectangular face? I say it doesn't, that it's a red herring, and have previously given reasons based on (a) the improbability of the cut off being at  much the same distance on both sides of the vertical midline (b) the bulge at the cheekbone which lies across the cut-off line and (c) the baked-in neck/chin crease (an integral part of the image in my opinion)  that is also indifferent to the region of image-interruption, i.e pale areas between cheek and hair.

If it's not banding, what is it that causes the cut-off? Well, if the facial image is rather too gaunt and mask-like, then that suggests, does it not, that it was derived from a gaunt and mask-like template, which in turn suggests, after Luigi Garlaschelli, that a bas-relief template was used?

Somehow, don't ask me how,  the thermal picture in ImageJ has a blindspot for banding - the face in 3D colour-coded mode no longer looking so rectangular, having a wavy edge, with scarcely a hint of linearity or banding in sight.

It may not be clinching evidence, but it's one more nail in the coffin for the facile "banding-explains-all" cop out.

Reply:

David Goulet commented on Banding? Is it Real?.

in response to Colin Berry:

If you can spare a minute, David, scroll down to the far end of my current posting re banding (revisited) and you will see your comment (2.31pm) and a possible clarification. http://colinb-sciencebuzz.blogspot.fr/2014/08/imagejs-thermal-lut-mode-for-generating.html

A good clarification of your side of the issue. I don’t really ‘buy’ one assumption though — that the face is too rectangular, too Gothic. If anything the face looks swollen to me, which is exactly what I’d expect of a death mask. I saw my share of corpses as the son of an undertaker, and the man on the Shroud sure looks like one. We are not dealing with a painting, we all agree on that. We are dealing with a yet to be confirmed image process. Because we do not know how it works, we cannot rule out any image distortion caused from that process.

Your research is in an effort to prove that a medieval scorch theory cannot be ruled out. You’ve made some strides there. On the other hand, nothing you have discovered, to me, has ruled out a 1st century origin. I believe you’ve acknowledged this as well, pointing out that a scorch (in the broader sense of the term) theory might support a 1st century origin.

Question: a bit off topic perhaps. Can sound waves create latent images on surfaces other than liquid ones? The banding (lines) on the TS remind me of sound ripples in some pics. Too sci-fi perhaps, but could an ultra-sonic pulse have created a reaction on the linen forming an image? A Big Bang mistaken for an earthquake perhaps? Sorry, the creative writer in me is bleeding onto the page at this point.

My response: The immediate goal for me is not to try proving or disproving that the TS is first century (for that we'll need repeat radiocarbon dating, if only to dismiss the re-weaving story), It's to demolish the various dogmas still doing the rounds about the TS image being different in character to any thermal imprint, medieval or otherwise, based on dated experiments and observations, ones that look increasingly redundant with each passing day.

Contact scorches can be highly superficial, at least at the thread level, and can respond just as well as the TS to 3D rendering etc etc. However, the chief reason for thinking the TS is a contact scorch is the negative character. I cannot think of any process except contact imprinting that can produce a negative image, considering the radiation theories to be frankly barmy - pseudo-science by any other name.

Thursday: finally, to draw a line under this over-long posting (a more detailed look at that "min slider control" will be the subject of my next posting), here are two comments that have just appeared on Pirate/Troll Central, the first from the new(ish) BSTS Editor, Hugh Farey, followed by my congratulatory response.

Hugh Farey

August 28, 2014 at 3:42 am

Reply

In Shroud 2.0, longitudinal banding is very clear, and is definitely related to the pitch of the zigzag, specifically the darkness of the shadows cast by the overlying warp threads onto the underlying weft threads. Thus the entire Shroud is covered in alternating lighter and darker bands. This pattern is not seen on the Durante photo. Here the various longitudinal stripes seem to me to be much thinner, where you can see them, and appear to be related to the ‘spines’ of the herringbone ribs, which may have formed into slight ridges or troughs as part of the rolling up process. I cannot find a good positive Enrie image, but the large scale negatives, which can be found at the link above among others, show a variety of bands, some very thin and some as thick as a width of a pitch. However they are much less consistent and the thick ones do not appear to be lighting artifacts as they sometimes extend over two or three bands of alternating pitch. It is not clear in any case that the pale vertical areas defning the sides of the cheeks, or the dark vertical areas defing the fall of the hair, are due to imperfections in the weave or the lighting of a photo rather than the shape of the image model itself. As such, attempts to ‘correct’ the image by removing them are probably misguided.

Colin Berry

August 28, 2014 at 4:09 am

Reply

Brilliant. Hugh. Possibly, nay probably the best contribution to ‘banding’ in all time.
Re my new toy (Thermal LUT in ImageJ), I have been playing around with the “min” slider control. I now believe it to be a powerful aid to TS image analysis, having figured what it’s doing using model 2D shade diagrams with no 3D history. It basically drills down bit by bit from the high points on a 3D rendering lopping off the peak value progressively as it goes, probing deeper and deeper into the lower levels of the relief. The ability to do this in degrees, comparing the TS with model at each stage, makes it, in my opinion, a valid and powerful research tool in image analysis.
I’ll post my model validation study today or tomorrow, and will then take a break for a while, leaving this vexatious site (but for you and David Goulet) for yet another cooling off period.