An elderly friend visited the palace a few years ago, even managed to take a pic with the queen up against the gates (directly behind them)…but finger fumbling failed and they accidentally set this artistic filter. So their precious memory is ruined
The original camera and memory card with the original file has been retained.
Is there even the slightest chance the original unfiltered file might just be on the camera? It would make his year if it was. May be there’s a function or process that can revert to the original?
Damn… that’s a tough one. The short answer would probably be no.
But it depends on a bunch of factors.
What’s the camera? What are the photo taking settings and does it save an original file next to the filtered one? More pro oriented cameras have an option of RAW + Jpeg or jpg. So if that’s there it’s worth exploring.
Thank you @MCVET , I thought I had details of the camera, but can’t find them at the moment. I have to go back to install a new SSD, so I’ll grab it then (in less than the next 2 weeks).
All I did so far was look around the settings for anything relevant, but didn’t have any joy. I never thought to remove the memory card and plug it into an external reader, with the hope there’s a hidden folder with originals.
He would be so overwhelmed if it existed, being 86 years old, there’s few pleasures between medication to keep things moving!
Thank you @anon7678104 , honesty is best even if it’s not good news. I’ll certainly be doing as you say, fingers crossed!
Thank you too @TimHolus , you never know, they might want an image to test their cleverness on!
You could setup their camera to shoot raw. Basically it will give the RAW data to the card. It won’t help with this situation but may be useful in the future.
I doubt it’d work, but in the slim chance that the camera first saves the original image, then reads it and applies the filter, then saves the filter version and deletes the original, then the original is still there, just unallocated in the FAT, in which case PhotoRec would be able to extract it (it doesn’t use the FAT).
Damn, that’s rough. Just as many said already there’s the chance that the RAW file might still be there or through the camera would be possible to use the same photo to recover the original file (like “apply no filter”). If the camera just shoots in JPG I’m very sorry to announce that there’s no way to recover it unfortunately.
The processing phase applies the filter before writing to the SD card so there hsa never been an original untouched photo on it.
Maybe you could use some ML softwares to create all the details, but that’s about it.
I understood that just before taking the photo or just after it was taken / in the process of saving.
However, this does not change the fact that the matter looks rather weak. If there is no original file on the camera before the filter was applied, it’s almost over.
Unless the OP finds someone who plays with very advanced image filtering based on algorithms and sample databases and will just recreate something that will look like an original photo.
Reconstructing a photo that has been subjected to a certain filter is possible to some extent, but it is not a walk on the beach. And I don’t know if this particular filter would even allow it. But theoretically it is possible with other filters…
If we know the characteristics of the filter, then we can know it and reverse it. For this you need to add data that is changing, which does not exist.
If we know the value of A (we do not know) and the process of changing B (not currently), and we have the final result of C, theoretically, something should be able to be rebuilt in the reverse process as long as the algorithm is sufficiently advanced.
If the final result shows the value of C123 and we know fully how the filter works, we should be able to guess what value could be contained in A***.
Assuming that the filter will always compose on the variable constant, it should be possible to guess what value could be included in the original if the final result is C123 and the variable always causes A375 to always get the resultant C123.
A lot depends on the filter algorithm itself and how it computes the variables relative to the original. If there is a fixed rule, theoretically there is light in the tunnel.
If the filter does anything randomly then you can forget about the matter.
You need to know the principle of the filter algorithm, i.e. see why the A345 pixel is changed to A567. If it is a constant and repeatable rule, it should be possible to calculate the opposite. If we have A567 and we know in order to get A567 the source had to be A345 then we can rebuild something close to the original photo.
Absolutely. Although RAM pics are larger in size aren’t they? I’ve got a feeling his days of taking pictures has come and gone now, so it shouldn’t be as much of an issue.
Fingers very much crossed that’s how it does it. Thank you for the photorec link, I’ll make a note and you never know, could be the photo saver!
Yeah, totally sucks for them. That’s sad about jpg.
The processing happening before writing to SD card, makes it all the less likely. We’ll see though, I’m going to pick up the camera on my next visit to him and see what I can do.
Yep, although I guess when out and about creating media, it makes it difficult - especially when the tech user is over 80 years old.
Sadly not - finger fumbling on the camera before taking the picture. Somehow changed from unfiltered to filtered
You are way cleverer than me
I do understand though, it’s as if we ultimately need a test shot to identify what it does with what context, making it easier to reverse the process. That is above my pay grade though, as they say in the movies! You never know though, someone might be able to do the algorithm reversal!
Here’s another pic…the King and Queen walked past them
Problem is, the camera user COULD have previewed the pictures taken, seen the error and very quickly reverted. But they didn’t…I don’t look forward to getting old and continuing to use tech, mistakes will happen!
If we take photo A1 and apply the filter. And again we will take the same A1 photo and apply the filter again. Will the pictures A2 and A3 be 100% identical in both cases? If so, then we operate fairly with a constant characteristic when it comes to a filter.
The question of the problem is how exactly the algorithm of this filter is created and how aggressively it calculates the changed values. A lot depends on how much data is lost in the modification process and whether random change events occur.
The problem is also the fact that even if we know the data reversal mechanism, we will still have the missing data in the already modified photo.
This would require a very precise algorithm that will reverse segment by segment understanding specific values and their original position. You can feed the AI with photos of this place as well as photos of the person you want to be visible and create layers that will be assigned to the coordinates in the photo.
Our algorithm had to understand what to expect in the original photo and understand the filter algorithm and then try to build something.
Because even if we know the filter’s operation scheme, we will still not be able to calculate the original value in many places of the photo. We will not calculate that the black pixel in the xyz position and we will not guess what was there earlier if we do not introduce a very precise pattern that will feed our filter about the probability of the value of this pixel in the original photo.
So our algorithm must understand the operation of the filter and understand approximately what is the state of the original photo and try to calculate the potential changes between state A and state C.
Smartphones really didn’t bother, with hosting .RAW file support, until recent years
The photo process being basically blackboxed, where the intra-files wouldn’t be kept
