It would be great if Topaz could allow us to give Sharpen AI a reference photo that would help it do a better job in sharpening. For example, I may have 2 very similar photos. My favorite one though is somewhat out of focus and/or has motion blur, but the other less interesting photo is sharp. Maybe a person who moves at just the wrong moment in the more interesting photo. It would be cool for Topaz to look at both the interesting unsharp photo and the uninteresting sharp photo in order to sharpen the unsharp photo.
I’m voting this up as a proposal.
1 Like
When recovering faces from images that are blurry/noisy/bad resolution, faces that are known to me often come out as a more generic face. Like an old picture of my kids don’t look as MY kids anymore, even though they are sharp and noise-free. Would it be possible to make the AI for faces go through MY locally stored images and make an additional locally stored training dataset, that the AI may use when dealing with my images. This would mean that each user could have a locally optimized AI.
I suggest this since I don’t think uploading my images to Topaz for general training would help, since it would “drown” in the general pool of data (millions of images) used to train the algorithms, so would not give enough “bias” towards recognizing the features of my subjects in particular.
Just an Idea, I don’t know if it’s even possible…
Regards
HAL
Late to see this post, but Stable Diffusion should be able to do what you want (after becoming familiar with the application and models available). Even without local model training, you can get fairly close to a good representation of the face of the person of interest, by use of the controls.
But you can also train your own model using your own photos e.g. using Dreambooth if you have a lot of them (or good quality video). This is all new stuff, free and time consuming to get to grips with but under continuous development and it’s just a matter of time for it to find its way into easily manageable software. Meanwhile, unless you are a DIY guy with a decent GPU, sit tight and wait…
Voting on this product feature request. Topaz Photo AI produces unacceptable superficial faces when enhancing low-resolution photos. TP AI should allow reference photos during the process to ensure that the correct facial features are maintained, and the “identity” of the person in the photo is not lost. That would improve the AI!
2 Likes
I’ve seen other posts that dance around this topic, but none specific to recovering faces.
With low resolution or blurry photos, Photo AI frequently produces facial results that do not match the actual person. It has no way to know this, so it would help to be able to submit a reference photo of the person or persons being recovered so Photo AI can “learn” what the face should actually look like.
In the “Recover Faces” section of the UI, there could be a button that asks “Faces don’t look right?” Pressing this button would open a modal and allow upload of another headshot of the person in question. This could help improve facial recognition and, over time, as Photo AI continues to learn, mean it’s no longer necessary to upload a reference photo of a specific person.
1 Like
I have also seen this request here:
https://community.topazlabs.com/t/face-training-for-known-faces/34851
And in both cases I have voted for it. This would be my number 1 enhancement request!
2 Likes
When attempting to improve face quality it would be immensely helpful if users could upload a folders of high res. photos for those who are in the videos.
If Topaz Video AI could rightly identify the faces in a scene and then weight the faces in the reference stills faces to use for detail I suspect the accuracy results would be much higher than trying to use generic facial feature up-rezzing (which will result in sharp, but in accurate – Uncanny Valley type faces.) Using actual faces of the subject, sourced from photos of the same time frames, ought to produce amazingly detailed faces HD scaling with the unique benefit of having the face appear most accurate.
1 Like
Wow, you’re asking for a lot! I think perhaps you’re a couple of decades ahead of where this is going someday.
What you are describing is basically how deep fakes work, were you can supply a bunch of source faces to learn from and it replaces the destination face. Deep fakes is a bit more general where you are applying a source face to any destination face, but it can be used to replace a low res face of the same person too. It would be interesting to see a combination of DeepFaceLab → Topaz VEAI and see if it helps with the temporal consistency of the face.
yes! What better way to uprez a face in a low res image than to use high res pictures of that very same face?
2 Likes
Allow the AI model to accept input images to finetune its training/model for enhancing photos of specific people.
I have a few decent old images of my grandparents. If the model could learn how they looked, maybe the final result would be better on the many lower quality images that I have?
EG: better refined facial features and more accurate shapes, smile, smile’s impact on the eyes, eye color, hair styles, facial hair, hair color, etc.. At the moment I can process 3 different LQ photos of the same person and the output can end up looking like 3 different people.
Hope it comes someday. Im searching for this in GitHub, but dont find something easy to do like: Put 10-20 HD images and let the program run to upscale a face.
3 Likes
That would be absolutely great
1 Like
This would be my number 1 enhancement request too!
3 Likes
+1 to this feature! It seems that if a photo is terribly out of focus, Photo AI struggles to restore the original facial identity of the subjects. The sharpened or regenerated face looks make the subject look like a different person. It would be great if I could upload 1 or more good reference photos to help Photo AI to reconstruct the out of focus subject. Kinda like how deep fake sort of works, I’m sure it’s a very possible feature to have.
1 Like
I posted this suggestion 3 years ago (!!!) for Sharpen AI and before Photo AI came out. Now, of course, I want it in Photo AI too.
2 Likes
Henry,
I was thinking about something similar. However, with TPAI, the reference photo could have multiple uses. Imagine it in noise reduction, sharpen, and remove.
Since the remove feature currently generates whole parts of an image (by guessing what would look okay there), think of how precise it could be if it had another image of the same place without the offending object or person, or simply from a different angle. It would then “know” what’s behind the object and how to fill.
As already mentioned, a blurry face (or whole person, or hand, etc.) could be easily fixed with a reference. You’ve taken an almost perfect beautiful soul-capturing portrait of someone who unfortunately moved his or arm right when the shutter actuated – I know you have!
…Sam
1 Like
Please, this is exactly what I’m looking for!
2 Likes
I just came here to request something similar…but not just for faces. When I’m out shooting photographs of wildlife, I’m often taking a series of photos where the action I hoped to preserve is in the one frame that happens to be either out of focus or has a bit too much motion blur. But often other frames of the same subject are tack sharp. If only one of those frames could help inform the AI on how the blurry frame could be repaired…
