New NGC slab type

Not so...today's laser scanners scan at a multiple angles to collect x,y, and z data (which is 3D). You see this in a number of industries including everything from the military to dental digitized scanning. That is why I mentioned x,y, and z data points (3 points), not just x and y. Laser scanning can collect data down to as small as 10 microns. That sure is better than how coins are graded with a 5X magnifier. Remember...the key is what is the given? My suggestion is to take 10 coins of a given year and mint mark from each of the top TPG's and make a collective database of what an average is of a given coin from a given year and from a given technical grade. Of course 10 microns is way too small, but you can always adjust how many datapoints you wish to measure.

Our eyes can see items very well. The problem is in processing and communicating color...hence the dot experiment I posted. Our eyes see color differantly under the sun than in your kitchen. The given item's color didn't change, but the lighting and adjacent colors did! For instance, you could NEVER take a picture of your car and then head to your body shop to buy paint that matches the "true" color of your car. That is why they use color measuring instrumentation that goes directly on the car to measure the color ACCURATELY. The same goes for places like Home Depot. Pick a color in their store and people wonder why the same color looks differant in their living room. Why? the lighting is differant.

Bottom line, there is NOT one industry that use photography to measure and quantify color. Is the lighting the same? Is the distance of the light the same? Is the intensity of the light source the same? Are the adjacent colors next to the subject the same? Is the angle of the light the same? The answers to these questions are all NO and hence the reason why photography is great for taking and showing pictures, but never a true exact representation of what color is it.

Don't believe me? Take your favorite coin and take a picture with the same camera outside and then take the same picture of the coin inside. Same coin, differant color results.

BTW...here is what x,y, and z data looks like. Imagine 50,000 plus data points per coin. It can happen...the question is would the TPG's want it to happen.

Speedy...the technology wasn't there back then as it is today. Laser and digital scanners weren't even around 10 years ago. Think about how much technology has entered the market place over the last 15 years alone. As for your conspiracy theory and being able to adjust data, I guess you don't trust a computer to add and subtract. Let's have a race and see who is more accurate. My excel spreadsheet column with 500 figures against your pencil and paper addition. Ready, set, go!...I am done.

Consistency, consistency...how do we achieve that? Here's a start.

"It's common sense to take an idea and try it. If it fails, admit it and move on to the next...but more importantly, try something." - F.D.R.

 

Log in or Sign up to hide this ad.

Speedy: blame the admins who aren't maintaining those machines properly. No excuses, they are getting paid for doing proper maintainences or the stores aren't looking for a proper admin. In reality, machines have to work all the time or else you will have a screwed economy. Taking the example of a stock exchange, if the main and backup machines go down, you will have businesses going down and amounting to millions if not billions of dolloars of loss. So far, this rarely happened or not to what I know of thankfully.

Midas: Yes I know those exist. But do you have an idea of how much cost to set up such equipments? Those equipments are mainly used in the medicial and geological surveying teams but those big players DO have the money to play around in the first place. Maybe if the grading companies have more than 100+ million dollars to play around with (that's what I heard from some mining surveying teams) (probably a lot cheaper when it comes down to coins), then this can be a reality. Perhaps, this might be an excellent tool to differentiate M69 and the ever "scarce" MS70.

But even so, this only solves modern coinage - it does not help to solve crude striking coinages, ie pre 1800s. No matter how big database it can ever get, I just don't see how it can cut it.

"OMG AWFUL PLANCHETS DURING THAT TIME!!! HOW TO GRADE?!?!?!?!"
"manually struck coins - uneven"
"overstriked coins, too many features!!!!!!"

Plenty of blue screen of death, mate.

Also, how can this be put to use against coins with slight dirt against old-proof / metal parasite damage / PVC / cleaned / whizzed, etc coins???

 

Ha ha ha As if anyone needs more of that.

 

I think the real problem is entropy. There is no real algorithm to determine what 'normal wear' is, because if you're talking about imaging at such a refined state that the softwear alone can determine grade designation with any accuracy, you're talking serious abnormalities that the software has to be programmed to assimilate. Then you can only allow for so much variation before the software can no longer determine between one grade from the next in sequence.

If you can get past that, which I think would be serious business, by itself, then you have the problems already mentioned. How can a computer software determine what is cleaned and what is circulated worn or corroded or otherwise intentionally damaged?

I would grant that computer imaging can be introduced to the process, if and when it is cost affordable, but I think we are a LONG way off from computers taking over the process entirely.

 

I remember reading somewhere that they tried (who?) computer grading about 10-12 years ago. Before digital photagraphy, and back when software could not handle the task. I still do not think computer grading could or should be viable. No computer grading can take "EYE APPEAL" into account.

 

I used to see a lot of the "computers will never be able to _______" type of arguments applied to computer chess as it pertained to mimicing human intelligence and judgement. Now world champions have a difficult time holding their own against the mechanical monsters. It's pointless to refer to what failed in the past, because technology is always advancing. It's pointless to fall back on esthetics, because programs now exist that can scan a group of human faces and predict accurately which one will be voted "most beautiful." It's pointless to refuse to rely on a human programming team, but rely on a human to grade a coin. Maybe someone will devote the research dollars to come up with computerized coin grading. Maybe it won't be considered economical, like inventing a lettuce-picking machine to replace illegal immigrants. But it most likely isn't beyond the ability of technology to mechanically grade coins accurately today, and almost certainly will be possible in the future if it is decided to be worthwhile and provide an economic return to someone.

 

bzcollektor
That would be COMPUGRADE---
http://www.sampleslabs.com/compugrade.html

Speedy

 

Cloudsweeper, of course you're right - technology is growing at a rapid pace, and I don't think anyone here would argue otherwise... Who would have guessed, back in the early 80s, just how much we would rely on computers today.

But where are we now?

I see problems with computer grading, and I don't see those problems being solved in the near future... eventually, of course they could, and probably will, but I don't see it any time soon.

 

I'm not sure why you keep repeating that. What is the incentive? How are you going to get a group to collaborate to deceive like that from day one? How will you fake the beta testing? How would you successfully launch the product? How would you get customers to pay for a service that is flawed from the start? What evidence is there that the program will operate inconsistently with its coding? Why is it assumed that a rigged program will somehow reward the developer more than an honest effort? What is the incentive for the programmer to spend a lot of time on something that may eventually send him/her to jail? Why do you find it necessary to trust only programs that were not developed by humans? Is this the way you think about all computerized activities? Do you believe ATMs are designed to cheat? Are supermarket price scanners programmed to cheat? Do all medical test facilities purposefully send back false positives? Nothing in that post makes any sense in the 21st century.

The comments about eye appeal I can understand. It remains to be proven. But to condemn the entire programming industry as incompetent and dishonest really isn't much of an argument. If it works, it works. If it doesn't, it needs more work. Surely there is someone out there interested in writing an honest program to solve an intersting problem. Programmers succeeded with chess computers by collaborating with chess masters. They will succeed by collaborating with reputable, qualified coin graders too. There is much more to gain in money and prestige than by going into a project that will take thousands of hours with the idea that the intent is deception and the goal is ultimate failure.

 

rick - If we don't see computerized coin grading, it will probably be because the economics of the project don't make it feasible. I don't think it will be due to technical problems. This isn't the toughest project in the world. I mean, we're not building a Mars rover or cruise missle here.

 

Well, I wouldn't underestimate the complexity of the solution.

How many variations of VF have you seen in your life in one year, one type coin?

 

Necessity breeds invention and this hobby is screaming for something consistent.

I also like what Disney stated many years ago, "If you can dream it, you can do it."

I have seen laser scanners capture 3D data right down to every nick and mark. I was in one company talking about our hobby and they asked me for a quarter. I pulled one from my pocket and they scanned it. They then showed me the data points on a computer screen. The data was obviously stored and then this data was sent to a computerized milling machine where they milled the reverse of my quarter...right down to every scatch and nick. Since they didn't have the right alloy to mill, what came out was the exact reverse of the quarter I just had in my pocket. BTW, they milled the material with 3 differant size burs that automatically changed throughout the process. I couldn't keep the "coin" because the last thing the company needed was the Secret Sevice knocking on their door, so it was destroyed.

I almost forgot. Total scan time. About 30 seconds!

They explained how each point on the 3D surface (the coin) was scanned for x, y, and z data. Since this company deals with dental restorations, they came up with a laser scanning digitizer that can scan a prepped tooth so that a milling machine can make for your dentist a new tooth (crown or cap) that fits perfectly over your remaining tooth structure. Now you can have a same day, same hour, new crown instead of waiting weeks for an outside lab to make it...which requires at least 2 appointments. Accuracy is a must so data is collected every 5 to 10 microns (we're talking small).

After explaining "our" problem, they told me the key would be developing a database where mostly everybody is satisfied that a MS66 is in fact a MS66. Color could be addressed with colorimeters and other devices. But for the most part, technical grades at the very least can be completed with the right software and programming. The hardware is HERE, we just have to figure out how to apply numismatics to it.

Costs for a laser scanner and digitizer...about $80,000.

"The successful person makes a habit of doing what the failing person doesn't like to do." - Thomas Edison

 

I see plenty of problems with the new technology. Wait till you get some mafia men getting hold of those machines and all you know next is that ultra-high quality counterfeits are made by them, all thanks to the brilliant suggestion. Metal alloy tests will fail next because they will melt down some crappy worn coins that were made around the same era, which is happening right now. 80,000dollars (maybe more if you include the die engraver) is dirt cheap to them if you can strike plenty of high quality fake coins that's worth 10,000+ dollars.

The last option we might have left is to check how much "dirt" there are on the coins but that will be another expensive process.