• Content Count

  • Joined

  • Last visited

Personal Information

  • Occupation
    my wife wishes she knew
  • Hobbies
    coding and collecting wine, art, money....
  • Location
    on the move... why stay in one place?

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. I purchased this 1740 British half penny recently. When it arrived in taking a look I said that the "0" looks a lot like a "9" Here is a 1740 but then I took a look at a 1730 and the "0" bleeds into looking like a "9" as well. Here is a good picture of a 1749 What do you experts think?
  2. got it. thanks. obverses are in even worse shape. uploading if interested.
  3. came by these 2 ancients today... just starting the research. anyone have any quick ideas? thanks.
  4. 1823 Mexico 8 escudos KM#382.2 NGC 1823 was the year Mexico transitioned from a Monarchy to a Republic.
  5. I don't know anything about acoustics but eager to learn how to apply it in this use case. Do you have any material to read on the subject? for those still interested in the camera aspect and remember the company Coinscope here is an interesting expired patent from them:
  6. identifying a coin is pretty easy using something like tensorflow and just scraping the web for 5 to 10k images, categorizing them and then training on them. like i said, i did that in a few hours (front of coin only) and have it up to about 80% accuracy if the photo uploaded is good. going the next step (date/mint) is a bit more difficult but doable. Face detection then recognition is a well studied field. here is some simple python that will explain the detection part using opencv once you detect you can narrow against another training set. Using ML to do the same thing with coin dates is straight forward as well-- it just takes time to build the training set (cutting up hundreds or thousands of images, then categorizing them). I just don't have the time or the patience to build that dataset. That's why I was looking for one of the DB's outlined in the second article and they are not easy to come by for download. Taking an easier road would be to do it with OCR and use Tesseract but I found that the font is a problem on most coins and it won't ID them correctly. You could train tesseract and I might try this in the future but I didn't go down that path.
  7. enjoy, if you want to continue to geek out--read this later these databases are hard to find but there was really good work done on this subject by some clever people a few years ago. shows how hard it is to crack the code on the date issue but it has been done before.
  8. but... if you have the right equipment (like an camera -- pcgs owns a few -- see this great pix ) and you have the images (like NGC/PCGS/HA - categorized by grade) and you are in possession of the coin so you are in control of the camera then this is pretty straight forward and any dev that has some deep experience with tensorflow or pytorch could build it.
  9. CRAWTOMATIC - yep, it has a lot to do with strategy and the goal of the system. Is the goal for a third party (dealer) to use it, is it for a grading company to use it as the final word or is it to help the grading company in quality control. if it's quality control, then overlaying a mask over it like the one attached below and grading only the anomalies in black can work and is pretty straight forward. I suspect that PCGS is already doing this given their early work on the subject. There have been a few articles published about their work they even did a video on using it for counterfeit a couple years ago. figuring out counterfeit if its a defect (misaligned or vampire bite marks) is really easy from photos. what material its made of -- not so much :-)
  10. More on the topic that many of the generic coin auctions are for suckers and are marketing gimmicks (HA not included--they are actually pretty good)... Here is an example: This company liveauctioneers has been auctioning off and supposedly selling this same 1879-cc morgan MS 63 for $8500, then $8000 and then $7500 (PCGS has it at $6850) since 2017 (maybe earlier). They list it in an auction and then the computer bids the audience up but ends up not selling it and the computer wins because the company didn't get their prices. Audience (if there is one), just thinks someone else purchased the asset at a high price. They also do it under multiple different Auctioneer brands (I wonder why :-) ) -- net, buyer beware.... Morgan Dollar&sort=-hammerPrice&status=archive
  11. CRAWTOMATIC - their patent was written at a time when many of the current techniques were not known. Things in tech have changed and made a lot of what's in there easier IF you have the coin in your possession (so you can control camera and number of images) and IF you build your trained model with enough high quality images per classification. Anyone can hire the tech to build the model if you have the images. It took me just a few hours to build the model for the identification on Tensorflow. I am stuck on the date issue as OCR doesn't work well. Even without building a trained model any 10th grader could build something that does an absolute diff on a threshold built from a bunch of high grade images (like in paper) as part of normalization prior to training. I put my 10th grader hat on and did it in a few hours but then spent the rest of the weekend stuck on the date/OCR issues and the glare in the photo issue (both difficult to normalize for...)
  12. This post probably cost me about 100 hours of fun :-) I spent a couple weekends trying to do this following this paper: --it's not all that difficult but the problem is that it requires a lot of high quality pictures per coin, per grade, per year, per quark (ie. 1909 VDOB) back and front--i estimate about 1M pix for US coin base all up. Then it requires the user of the system to take the same high quality photos. There are big issue with normalizing glare of high end coins. net is only a HA, or TPG has the base of photos. It's pretty easy to take 1 photo and munge it up into many but the score goes down significantly. Biggest issue I had is with identifying date--it's a much more difficult problem than identifying a license plate. Note that PCGS has been doing ML "assisted" for a while "With PCGS Gold Shield, each coin is imaged in high resolution, registered, and checked against PCGS’ vast proprietary imaging database. Our Gold Shield system benefits from artificial intelligence and machine learning, helping our graders quickly identify and remove counterfeits from the market." Interesting enough they let their patent expire: I'll spend a couple more weekends on it as it's an interesting rabbit hole for a geek that likes coins...
  13. sure, dealers use CDN to buy, but they use PCGS/NGC prices in marketing the asset. Stack’s Bowers links right to the pricing page LiveAuctioneer does it David Lawrence does it eBay does it even the small auctions in far off places do it and heck, PCGS even references NGC values
  14. VKurtB - yep-big world... room for everyone... even those with a disjunctive union :-) I don't disagree on the groups not being good at valuation... saying something is worth 55k or even 24k when it's really worth 14k shows a pure lack of proficiency at the task. "Nobody takes them as gospel" -- I'll remind all the coin dealers :-)