A different type of overlay
messydesk
Posts: 20,007 ✭✭✭✭✭
I did this photo overlay exercise ATS with the Shield nickel IIa, IIb, and IIc reverse hubs. While it's not what you're used to seeing in a photo overlay, it produces an interesting result where you can still correlate each coin with its features. Step-by-step, assuming a basic understanding of using Photoshop layers, here's the process:
1. Put each picture in a different layer and make a traditional overlay, making rotation, translation, and scaling adjustments to line up fixed reference points, in this case, the 5. This shows there is a difference, but it makes it hard to tell what they are.
2. Run a high-pass filter on each layer, with a radius of 3 pixels. (Only one layer shown). This accentuates the edges and makes the background of each layer (i.e., the blank fields) all the same gray level.
3. Adjust the curves on each layer to make the fields (currently gray) black and the edges (currently black and white) white. The curve adjustment looks kind of strange, but this does the trick. (Note: while the edge detection filter could be used to replace this and the previous step, the results don't look as nice.)
4. Apply the photo filter (under Image->Adjustments) to each layer using distinct colors. Since I had three layers to merge, I chose red (IIc hub), green (IIa hub), and blue (IIb hub). For two layers, green and red work nicely. Set the density of the filter to 100% and uncheck "Preserve Luminosity".
5. Set the blend mode for the layers to "Lighten." The result will look like this:
The 5 is essentially white. Severe inconsistency in punch positions seem to be between 11:00 and 2:00, where you see all three colors distinctly. Magenta (red+blue) is where IIb and IIc line up well (TED S). Cyan is where IIa and IIb line up well (left denticles, UNI). Yellow is where IIa and IIc line up well (parts of some stars). You can also see the ring of red denticle edges inside the cyan denticles on the left side, showing how the IIc denticles are shorter than the other two. There may be some error in the registration step, and by using points on the 5 as reference points, this misregistration will be magnified toward the edges. I think I got it pretty good, though.
1. Put each picture in a different layer and make a traditional overlay, making rotation, translation, and scaling adjustments to line up fixed reference points, in this case, the 5. This shows there is a difference, but it makes it hard to tell what they are.
2. Run a high-pass filter on each layer, with a radius of 3 pixels. (Only one layer shown). This accentuates the edges and makes the background of each layer (i.e., the blank fields) all the same gray level.
3. Adjust the curves on each layer to make the fields (currently gray) black and the edges (currently black and white) white. The curve adjustment looks kind of strange, but this does the trick. (Note: while the edge detection filter could be used to replace this and the previous step, the results don't look as nice.)
4. Apply the photo filter (under Image->Adjustments) to each layer using distinct colors. Since I had three layers to merge, I chose red (IIc hub), green (IIa hub), and blue (IIb hub). For two layers, green and red work nicely. Set the density of the filter to 100% and uncheck "Preserve Luminosity".
5. Set the blend mode for the layers to "Lighten." The result will look like this:
The 5 is essentially white. Severe inconsistency in punch positions seem to be between 11:00 and 2:00, where you see all three colors distinctly. Magenta (red+blue) is where IIb and IIc line up well (TED S). Cyan is where IIa and IIb line up well (left denticles, UNI). Yellow is where IIa and IIc line up well (parts of some stars). You can also see the ring of red denticle edges inside the cyan denticles on the left side, showing how the IIc denticles are shorter than the other two. There may be some error in the registration step, and by using points on the 5 as reference points, this misregistration will be magnified toward the edges. I think I got it pretty good, though.
John
Keeper of the VAM Catalog • Professional Coin Imaging • Prime Number Set • World Coins in Early America • British Trade Dollars • Variety Attribution
Keeper of the VAM Catalog • Professional Coin Imaging • Prime Number Set • World Coins in Early America • British Trade Dollars • Variety Attribution
0
Comments
i love that color overlay...
I like it!
The next time I want to compare dies, I'm going to try this method.
JH
Proof Buffalo Registry Set
Capped Bust Quarters Registry Set
Proof Walking Liberty Halves Registry Set
nicely done
really love creative thinking and there have been some threads lately that show just that, this one being no exception!
.
<--- look what's behind the mask! - cool link 1/NO ~ 2/NNP ~ 3/NNC ~ 4/CF ~ 5/PG ~ 6/Cert ~ 7/NGC 7a/NGC pop~ 8/NGCF ~ 9/HA archives ~ 10/PM ~ 11/NM ~ 12/ANACS cert ~ 13/ANACS pop - report fakes 1/ACEF ~ report fakes/thefts 1/NCIS - Numi-Classes SS ~ Bass ~ Transcribed Docs NNP - clashed coins - error training - V V mm styles -