I took a ton of pictures at an event. Unfortunately the autofocus of my cam is a bit wonky and this lead to blurry pictures. Does anyone have an app or way to restore details on the unfocused object? I also have the raw files of these pictures if that helps.
You can’t ‘restore’ those details because the camera didn’t capture them. The light was not focused on the same plane as the sensor
The best you can do is use an AI sharpening tool which will essentially guess what a sharp image would look like
https://duckduckgo.com/?q=ai+image+sharpening
From one of those sites: “The Focus model reduces lens blur caused by missed focus. This is the toughest use case and won’t recover 100% of your photos. When it works, though, it works really well and can completely recover a throwaway image.”
Your success will be highly dependant on how out of focus the images are. Slightly out of focus? You’ll probably be able to get it looking mostly right minus some small details. But very out of focus? The computer will never know what was there, because the camera never actually captured what it was.
Topaz Labs Sharpen AI is paid software, but does have the ability to specifically address out of focus photos. It’s hit or miss but can make a difference. I think you can try it for free, it’s worth a shot.
I’ll try that one out. If it works well enough I’m willing to pay.
Let me know if you need the torrent. It’s a great one.
Thx, but it seems to be a one time payment. If it where a subscription based model the torrent would be needed.
Cheers. :)
You don’t have a lot of options outside of an an unsharp mask or your standard sharpening tools. There are ai tools that are starting to come out that will basically rebuild your photo, but your YMMV and they require a certain level of technical know how.
Do you have any recommendations regarding the AI tools? Technical experience should be sufficient to run and configure tools when provided. Would be nice if there was a tool to sharpen a face based on other source pictures provided.
Take a look at ChaiNNer and it’s model wiki. Just a tip, you may need to downscale your image and then upscale it to get the sharpening to kick in. A lot of these models are based off of 512x512 sample images.
If you want to spend money, Topaz Labs might have something to offer, but I primarily deal worth video so im not super familiar with their still image offerings.
Will have a look at ChaiNNer in case topaz sharpen ai doesn’t get the results I want.
You will not be able to get the actual detail back unfortunately… However there are some of the ai tools that give best guesses as to what should have been there and like a sort of auto-generated picture can look extremely plausible. I have seen outputs from them and think the latest Google pixel phone will have it built in, but not sure were to access it online
Yeah I played around with different tools yesterday. Topaz sharpen AI was pretty much useless for what I need. But ChaiNNer with different models is looking promising so far to save at least some of the pictures I have.