Coin database

[STEVE MOULDING]

(this follows from the side discussion in squirrels thread..."Some finds at a local coin show")

 

Regarding my coin database...this is something I presented last year at a Numismatic Club:

 

http://www.russiannumismaticsociety.org/do...oinDatabase.pdf

 

However, it's already out of date (now I'm up to 400,000+ images) but shows the direction I've been working in.

 

Steve

[gxseries]

My eyes just boggled over the figure of 800 overstruck coins

 

And what a coincedence, I just sent Bobh about this issue. Just an off topic question though, how do you note the varities or do you keep a note elsewhere?

 

Steve, if I can help you in any way, please let me know. And 20 gb... that's what my previous hard disk space used to be.

[STEVE MOULDING]

Thanks very much for the offer of help! At some point I may need it, though it can be slow, tedious work. It takes most of my time trying to keep up with 18thC copper. I have 10's of thousands of images of Silver and 19thC copper waiting to be organized, and auction catalogs waiting for the scanner and/or photoshop cut & paste.

 

It's really a huge project (which I may never complete) and I've been at it for 5 years now. Unfortunately there's so much still to do so I can't always respond as fully or as quickly as I'd like if somebody asks me to check for a certain die/year of a Rouble (though there's a much better chance if the request is about 18thC copper). It has alreay resulted in some rarity articles for the JRNS however.

 

Die varieties I haven't even thought about. Still working on getting coins into their appropriate year/denomination folders.

 

On a different note, it would be somewhat tricky turning this into a public project / database as auction catalogs are copyrighted and part of the business revenues of our dealer friends comes from selling catalogs. I don't want to step on any toes there. Even internet images are likely copyrighted.

So long it's for personal research I would think its OK, but distributing it en masse probably isn't.

 

Steve

 

My eyes just boggled over the figure of 800 overstruck coins

 

And what a coincedence, I just sent Bobh about this issue. Just an off topic question though, how do you note the varities or do you keep a note elsewhere?

 

Steve, if I can help you in any way, please let me know. And 20 gb... that's what my previous hard disk space used to be.

[bobh]

Thanks very much for the offer of help! At some point I may need it, though it can be slow, tedious work. It takes most of my time trying to keep up with 18thC copper. I have 10's of thousands of images of Silver and 19thC copper waiting to be organized, and auction catalogs waiting for the scanner and/or photoshop cut & paste.

 

It's really a huge project (which I may never complete) and I've been at it for 5 years now. Unfortunately there's so much still to do so I can't always respond as fully or as quickly as I'd like if somebody asks me to check for a certain die/year of a Rouble (though there's a much better chance if the request is about 18thC copper). It has alreay resulted in some rarity articles for the JRNS however.

 

Die varieties I haven't even thought about. Still working on getting coins into their appropriate year/denomination folders.

 

On a different note, it would be somewhat tricky turning this into a public project / database as auction catalogs are copyrighted and part of the business revenues of our dealer friends comes from selling catalogs. I don't want to step on any toes there. Even internet images are likely copyrighted.

So long it's for personal research I would think its OK, but distributing it en masse probably isn't.

 

Steve

Steve, I read your interesting PDF file presentation about the database. It would be interesting to know more technical details about how you are going about things. For example, there is software out there -- some (a lot of it, actually) is open source and freely available -- which can run batch jobs on a folder of hi-res images, for example, and create thumbnail images or resize the existing ones, all without manual labor in Photoshop or your favorite paint program. That could save lots of time if you need thumbnails, for example on a web site or in some other graphical presentation of the data. Also, if there is a well defined folder structure and naming conventions, there are editors which can create HTML pages by programming macros to do the job (UltraEdit for Windows or Emacs on Linux, for example). Some databases can store images directly in a table; but some of those which can, aren't terribly good at it, so it might be better to leave the actual image files in the file system folders.

 

As to copyright issues for scans from printed catalogs, there probably aren't that many different venues to consult ... perhaps less than 20? As long as credit is given, there shouldn't be any problem as long as only images are depicted (for example, no prices realized, etc.) Besides, if the database were available as a public service and not for profit, you are providing dealers with free publicity by including their auction images.

 

However, if the database is publicly available on the internet, there is always the chance that someone will try to use some of the images for fraudulent eBay auctions. There's really not much you can do to prevent that except to discover the bad auction and expose it by reporting to eBay or whatever venue.

 

I think it is important to think about the scope of the application, whether it is strictly for private use, or available as a commercial or non-commercial public service. That will determine the direction any further development should take.

[grivna1726]

As to copyright issues for scans from printed catalogs, there probably aren't that many different venues to consult ... perhaps less than 20? As long as credit is given, there shouldn't be any problem as long as only images are depicted (for example, no prices realized, etc.)....

 

The fact that http://www.coinarchives.com/ exists suggests to me that probably no-one will get bent out of shape over the inclusion of catalog descriptions and prices realized as well.

 

I think it is important to think about the scope of the application, whether it is strictly for private use, or available as a commercial or non-commercial public service. That will determine the direction any further development should take.

 

The site I mentioned above is a non-commercial public service and my guess is that the dealers probably view it, as you suggest, as free publicity and a potential source of new customers.

[STEVE MOULDING]

(having problems with my editor)

Thanks very much everyone for the detailed comments and suggestions I'll be back with a detailed reply shortly.

 

Steve

[gxseries]

And I was thinking of the same lines as Grivna

 

Bob, along the lines of "However, if the database is publicly available on the internet, there is always the chance that someone will try to use some of the images for fraudulent eBay auctions. There's really not much you can do to prevent that except to discover the bad auction and expose it by reporting to eBay or whatever venue."

 

I believe the converse is actually more effective, i.e. it's a lot easier to report saying that "hey ebay look at this auction, it's stolen from this site which you can see at this link etc" rather than "available in ABC catalogue 10, number 100" which makes their jobs a lot easier in believing you and faster time to remove counterfeit coins as well. (if we push the issue of copyrights to one side for the moment...) But don't major auction sites such as Goldberg still leave their pictures online?

 

I really don't know how people view this but there are merits to limited access of free articles. I only found out RNS after I googled up Steve's site for the 1730 ruble that I have, as well as how I got my first contact with, and even more grateful that he shared his report on the Catherine II 1 kopek, which I swear if I didn't read it, I would have sold the whole lot without realizing a potentially interesting coin. It's actually much harder to find good infomation especially in English - I mean, you can google up Russian coins and see where RNS appears - not in the first few pages but it appears at the top if I google it up as Russian Numismatics. Counterfeit coins? I think we are at the stage of losing as newbies just buy them and these counterfeits have more resources to counterfeit them in a higher quality. I think a database of countefeit coins would prove to be interesting as well as hurting the counterfeiters in a hard way. The only way I can see to fight these counterfeiters is by through education which I got stung hard but I learned it well.

 

Forums such as this is really excellent but I think I have come to the stage that I have become too hungry for infomation and what has been bogging me down is the lack of Russian as well as my rusty programming skills which would have been very helpful

[STEVE MOULDING]

:

I second GX's offer of assistance. If there is anyway a volunteer could work on that project online, remotely, i can contribute some time. I work full time, but i have rather basic photoshop skills, and extremely slow typing skills to contribute.

 

Thanks very much everyone for your suggestions, and especially to Squirrel and GX for your offers of help (and to bobh who I've replied to privately).

 

A major piece of work to be done is the extraction/creation of individual coin jpg's from catalog scans. I have reasonably complete catalog sets for several dealers, catalogs that I've purchased directly from the dealer or have received from them as a bidder. World-Wide for example (Jim Elmen) I have all 52 (26 years!). While I've scanned 80% of his Russia-related pages (haven't got round to the more recent catalogs yet) I have only scratched the surface in terms of cutting out images with photoshop (or equivalent). My scanning/cutting is often driven by the particular JRNS article I'm writing.

Silver, Siberia, anything 18thC etc haven't been touched in terms of catalog extraction (though I have large sets of images from internet downloads).

 

If anybody would like to help I would suggest picking an area that interests them (which can be as broad as the 18thC, or as narrow as the 1718 Denga's of Peter I) and extracting all the images in that area from a single catalog scan (which I'd provide). If several people are working, it would take a little coordination but could easily be done. If the plan is to work on Jim's catalogs for example, I'd first talk to him and ask permission, with the understanding that this is not a commercial venture nor is something that will be easily available on the web. I foresee a research database, segments of which can be sent to interested parties by request.

 

Almost all of my work so far has been Copper of Peter III, Catherine II, and Paul I. Anything else is open.

Please send me a Personal Message if you want to help in this way, or have other suggestions. If not that's perefectly fine too.

 

Thanks again everyone.

 

Steve

[squirrel]

[bobh]

If anybody would like to help I would suggest picking an area that interests them (which can be as broad as the 18thC, or as narrow as the 1718 Denga's of Peter I) and extracting all the images in that area from a single catalog scan (which I'd provide). If several people are working, it would take a little coordination but could easily be done.

My former work as a software developer was not only in database work, but also document imaging. Although now I don't have access to the high-priced tools that I worked with then (2000-2004), I know that such things can be automated to a great extent once there is some kind of rules-based workflow in place.

 

I assume that all of your scans are JPEG grayscale images? I have some of Jim Elmen's catalogs here (thanks to you! ) -- what resolution do you use? Maybe I can come up with a script or a small console-based program that would save a lot of time for everyone (I do C and C++, too -- ImageMagick (open source SW) has a C API which could probably be used for this, although I'm not sure). Assuming that rectangular or square images are satisfactory, it should be easy enough to write something that would cover 90% of the cases. The other 10% could then be done by hand.

[gxseries]

Bobh, just took a quick look at the outline of what ImageMagick is, is it some kind of image editing software? I'll download that and give that a try. There is another image editing software that I really favour, (freeware by the way) that is Irfanview because of it's lightweight as well as it can crop, rotate, contrast, etc. http://www.irfanview.com/ I don't quite see what else it can't do other than cropping background. That's the only time when I switch over to photoshop although I detest how much ram it loves to suck up.

 

I'll leave the database programming to you

[STEVE MOULDING]

Bobh, just took a quick look at the outline of what ImageMagick is, is it some kind of image editing software? I'll download that and give that a try. There is another image editing software that I really favour, (freeware by the way) that is Irfanview because of it's lightweight as well as it can crop, rotate, contrast, etc. http://www.irfanview.com/ I don't quite see what else it can't do other than cropping background. That's the only time when I switch over to photoshop although I detest how much ram it loves to suck up.

 

I'll leave the database programming to you

 

I have seen irfanview and it is quite nice. I actually don't use Photoshop myself, but rather an image editing package that came with my scanner (PhotoEditor by ArcSoft I think). I'm quite happy with it. I just use the term 'Photoshop' to mean generic image editor.

 

Bob, yes you're right...my Elmen images are Greyscales. I scan at 300 DPI (fine for a database) but find that sometimes I have to turn a descreening filter on, else I get Moire-type patterns on the scan. I may be imagining it, but descreening tends to blur the image ever so slightly. Actually, now I think about it that may be exactly what a descreener does...some adjoing pixel merge algorithm which will obviously blur things. I'd agree that 90% of Jims lots could be cut out as rectangles, but 10% couldn't..obverse and reverse are in very different places on the page, or else a rectanglular cut would also take with it a chunk of some neighbouring coin). However, some auto-cut software sounds intriguing! It sounds tricky though.

 

Steve

[bobh]

I have seen irfanview and it is quite nice. I actually don't use Photoshop myself, but rather an image editing package that came with my scanner (PhotoEditor by ArcSoft I think). I'm quite happy with it. I just use the term 'Photoshop' to mean generic image editor.

 

Bob, yes you're right...my Elmen images are Greyscales. I scan at 300 DPI (fine for a database) but find that sometimes I have to turn a descreening filter on, else I get Moire-type patterns on the scan. I may be imagining it, but descreening tends to blur the image ever so slightly. Actually, now I think about it that may be exactly what a descreener does...some adjoing pixel merge algorithm which will obviously blur things. I'd agree that 90% of Jims lots could be cut out as rectangles, but 10% couldn't..obverse and reverse are in very different places on the page, or else a rectanglular cut would also take with it a chunk of some neighbouring coin). However, some auto-cut software sounds intriguing! It sounds tricky though.

 

Steve

You might want to check out this link:

 

http://www.siox.org/

 

I never used it before, but it looks like it might be able to do the trick, i.e. extract coin images from larger scanned images automatically (in spite of the caveats about greyscale images).

 

I found this link here: http://freshmeat.net/browse/893/. There are lots of interesting projects concerning image recognition which are open source, run on multiple OS platforms, and can be downloaded and used for free.