Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
Il giorno dom 6 ott 2019 alle ore 18:36 Aurélien Pierre ha scritto: > > Ok, let's go for a "sharpness score" + focus peaking feature. That way, users > will have a way to check which area is in focus, and compare images between > them with a score. > > It won't made it to next release, but maybe first minor update. Thank you so much! ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
I would love to have that feature. I find the current focus preview functionality useless. Br Jørn On 06/10/2019 18:36, Aurélien Pierre wrote: Ok, let's go for a "sharpness score" + focus peaking feature. That way, users will have a way to check which area is in focus, and compare images between them with a score. It won't made it to next release, but maybe first minor update. Le 06/10/2019 à 17:05, Robert Krawitz a écrit : On Sun, 6 Oct 2019 16:40:37 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: argh. Tales of over-engineering… I don't really disagree with you, just want to point out that getting it anywhere near correct (i. e. without a huge number of false positives and false negatives) is a difficult problem. Just overlay the euclidean norm of the 2D laplacian on top of the pictures (some cameras call that focus-peaking), and let the photographer eyeball them. That will do for subjects at large aperture, when the subject is supposed to pop out of the background. For small apertures, the L2 norm will do a fair job. And it's a Saturday afternoon job, hence a very realistic project given our current resources. That's fair, I just think that this kind of algorithm will likely select a lot of photos that are badly out of focus (because the focus locked on a much more expansive background) and miss ones where it's the relatively small subject that's in focus. What you ask for is AI, it's a big project for a specialist, and it's almost sure we will never make it work reliably. The drawback of AIs, even when they work, is they fail inconsistently and need to be double-checked anyway. So, better give users meaningful scopes and let them take their responsibility, rather than rely on witchcraft that works only in Nvidia's papers on carefully curated samples. Or maybe just implement focus peaking, as you say, but with a UI similar to the camera's UI (flashing regions that are in best focus). Then it's up to the user to select the best photos based on their knowledge of the desired subject. Le 06/10/2019 à 16:18, Robert Krawitz a écrit : On Sun, 6 Oct 2019 15:02:39 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: That can be easily done by computing the L2 norm of the laplacian of the pictures, or the L2 norm of the first level of wavelets decomposition (which is used in the focus preview), and taking the maximum. As usual, it will be more work to wire the UI to the functionality than writing the core image processing. Consider the case where the AF locks onto the background. This will likely result in a very large fraction of the image being in focus, but this will be exactly the wrong photo to select. Perhaps center-weighting, luminosity-weighting (if an assumption is made that the desired subject is usually brighter than the background, but not extremely light), skin tone recognition (with all of the attendant problems of what constitutes "skin tone"), and face recognition would have to feed into it. ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
Ok, let's go for a "sharpness score" + focus peaking feature. That way, users will have a way to check which area is in focus, and compare images between them with a score. It won't made it to next release, but maybe first minor update. Le 06/10/2019 à 17:05, Robert Krawitz a écrit : > On Sun, 6 Oct 2019 16:40:37 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: >> argh. Tales of over-engineering… > I don't really disagree with you, just want to point out that getting > it anywhere near correct (i. e. without a huge number of false > positives and false negatives) is a difficult problem. > >> Just overlay the euclidean norm of the 2D laplacian on top of the >> pictures (some cameras call that focus-peaking), and let the >> photographer eyeball them. That will do for subjects at large aperture, >> when the subject is supposed to pop out of the background. For small >> apertures, the L2 norm will do a fair job. And it's a Saturday afternoon >> job, hence a very realistic project given our current resources. > That's fair, I just think that this kind of algorithm will likely > select a lot of photos that are badly out of focus (because the focus > locked on a much more expansive background) and miss ones where it's > the relatively small subject that's in focus. > >> What you ask for is AI, it's a big project for a specialist, and it's >> almost sure we will never make it work reliably. The drawback of AIs, >> even when they work, is they fail inconsistently and need to be >> double-checked anyway. >> >> So, better give users meaningful scopes and let them take their >> responsibility, rather than rely on witchcraft that works only in >> Nvidia's papers on carefully curated samples. > Or maybe just implement focus peaking, as you say, but with a UI > similar to the camera's UI (flashing regions that are in best focus). > Then it's up to the user to select the best photos based on their > knowledge of the desired subject. > >> Le 06/10/2019 à 16:18, Robert Krawitz a écrit : >>> On Sun, 6 Oct 2019 15:02:39 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: That can be easily done by computing the L2 norm of the laplacian of the pictures, or the L2 norm of the first level of wavelets decomposition (which is used in the focus preview), and taking the maximum. As usual, it will be more work to wire the UI to the functionality than writing the core image processing. >>> Consider the case where the AF locks onto the background. This will >>> likely result in a very large fraction of the image being in focus, >>> but this will be exactly the wrong photo to select. >>> >>> Perhaps center-weighting, luminosity-weighting (if an assumption is >>> made that the desired subject is usually brighter than the background, >>> but not extremely light), skin tone recognition (with all of the >>> attendant problems of what constitutes "skin tone"), and face >>> recognition would have to feed into it. ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
On Sun, 6 Oct 2019 16:40:37 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: > argh. Tales of over-engineering… I don't really disagree with you, just want to point out that getting it anywhere near correct (i. e. without a huge number of false positives and false negatives) is a difficult problem. > Just overlay the euclidean norm of the 2D laplacian on top of the > pictures (some cameras call that focus-peaking), and let the > photographer eyeball them. That will do for subjects at large aperture, > when the subject is supposed to pop out of the background. For small > apertures, the L2 norm will do a fair job. And it's a Saturday afternoon > job, hence a very realistic project given our current resources. That's fair, I just think that this kind of algorithm will likely select a lot of photos that are badly out of focus (because the focus locked on a much more expansive background) and miss ones where it's the relatively small subject that's in focus. > What you ask for is AI, it's a big project for a specialist, and it's > almost sure we will never make it work reliably. The drawback of AIs, > even when they work, is they fail inconsistently and need to be > double-checked anyway. > > So, better give users meaningful scopes and let them take their > responsibility, rather than rely on witchcraft that works only in > Nvidia's papers on carefully curated samples. Or maybe just implement focus peaking, as you say, but with a UI similar to the camera's UI (flashing regions that are in best focus). Then it's up to the user to select the best photos based on their knowledge of the desired subject. > Le 06/10/2019 à 16:18, Robert Krawitz a écrit : >> On Sun, 6 Oct 2019 15:02:39 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: >>> That can be easily done by computing the L2 norm of the laplacian of the >>> pictures, or the L2 norm of the first level of wavelets decomposition >>> (which is used in the focus preview), and taking the maximum. >>> >>> As usual, it will be more work to wire the UI to the functionality than >>> writing the core image processing. >> Consider the case where the AF locks onto the background. This will >> likely result in a very large fraction of the image being in focus, >> but this will be exactly the wrong photo to select. >> >> Perhaps center-weighting, luminosity-weighting (if an assumption is >> made that the desired subject is usually brighter than the background, >> but not extremely light), skin tone recognition (with all of the >> attendant problems of what constitutes "skin tone"), and face >> recognition would have to feed into it. -- Robert Krawitz *** MIT Engineers A Proud Tradition http://mitathletics.com *** Member of the League for Programming Freedom -- http://ProgFree.org Project lead for Gutenprint --http://gimp-print.sourceforge.net "Linux doesn't dictate how I work, I dictate how Linux works." --Eric Crampton ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
Hi guys What would be really nice would be a nice tool to let the user choose, since from what I read, even with an IA we would never have, it cannot be reliable. I remember when I was using Aftershot (I think it was the name, before it was Bibble something I think), they had a REALLY great comparison tool : you select up to 6 photos in your list, then it displays them together (all next to each other), and when you zoom on one, it instantly zoom on all the others, at the same location. For a burst, it's great, you can very quickly tell which one is fine and which one is not. If I want to do that with Darktable, of course it's not impossible, but it's significantly slower. A similar tool would be great. Maybe without applying all the modules, we don't need a fully edited photo for the sharpness comparison. François Le dim. 6 oct. 2019 à 16:41, Aurélien Pierre a écrit : > argh. Tales of over-engineering… > > Just overlay the euclidean norm of the 2D laplacian on top of the pictures > (some cameras call that focus-peaking), and let the photographer eyeball > them. That will do for subjects at large aperture, when the subject is > supposed to pop out of the background. For small apertures, the L2 norm > will do a fair job. And it's a Saturday afternoon job, hence a very > realistic project given our current resources. > > What you ask for is AI, it's a big project for a specialist, and it's > almost sure we will never make it work reliably. The drawback of AIs, even > when they work, is they fail inconsistently and need to be double-checked > anyway. > > So, better give users meaningful scopes and let them take their > responsibility, rather than rely on witchcraft that works only in Nvidia's > papers on carefully curated samples. > Le 06/10/2019 à 16:18, Robert Krawitz a écrit : > > On Sun, 6 Oct 2019 15:02:39 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: > > That can be easily done by computing the L2 norm of the laplacian of the > pictures, or the L2 norm of the first level of wavelets decomposition > (which is used in the focus preview), and taking the maximum. > > As usual, it will be more work to wire the UI to the functionality than > writing the core image processing. > > > Consider the case where the AF locks onto the background. This will > likely result in a very large fraction of the image being in focus, > but this will be exactly the wrong photo to select. > > Perhaps center-weighting, luminosity-weighting (if an assumption is > made that the desired subject is usually brighter than the background, > but not extremely light), skin tone recognition (with all of the > attendant problems of what constitutes "skin tone"), and face > recognition would have to feed into it. > > > Le 06/10/2019 à 14:14, Germano Massullo a écrit : > > Il giorno dom 6 ott 2019 alle ore 13:32 Moritz Mœller > ha scritto: > > Define 'most focused'. > I give you an example to understand this request better. [...] > > > Yes you are right. but in your case, the couple is the main thing that > is moving in the picture. For my use case imagine I am taking photos > to people that are giving a talk. Some photos of the burst may be > blurred because I moved the camera while shooting, instead some other > shoots of the same burst could have less blur effect beause my hands > were not moving during its exposure time so the photo will have less > blur effect. > It would be great if an algoritm could detect the best shots > > > ___ > darktable developer mailing list to unsubscribe send a mail to > darktable-dev+unsubscr...@lists.darktable.org > ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
argh. Tales of over-engineering… Just overlay the euclidean norm of the 2D laplacian on top of the pictures (some cameras call that focus-peaking), and let the photographer eyeball them. That will do for subjects at large aperture, when the subject is supposed to pop out of the background. For small apertures, the L2 norm will do a fair job. And it's a Saturday afternoon job, hence a very realistic project given our current resources. What you ask for is AI, it's a big project for a specialist, and it's almost sure we will never make it work reliably. The drawback of AIs, even when they work, is they fail inconsistently and need to be double-checked anyway. So, better give users meaningful scopes and let them take their responsibility, rather than rely on witchcraft that works only in Nvidia's papers on carefully curated samples. Le 06/10/2019 à 16:18, Robert Krawitz a écrit : > On Sun, 6 Oct 2019 15:02:39 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: >> That can be easily done by computing the L2 norm of the laplacian of the >> pictures, or the L2 norm of the first level of wavelets decomposition >> (which is used in the focus preview), and taking the maximum. >> >> As usual, it will be more work to wire the UI to the functionality than >> writing the core image processing. > Consider the case where the AF locks onto the background. This will > likely result in a very large fraction of the image being in focus, > but this will be exactly the wrong photo to select. > > Perhaps center-weighting, luminosity-weighting (if an assumption is > made that the desired subject is usually brighter than the background, > but not extremely light), skin tone recognition (with all of the > attendant problems of what constitutes "skin tone"), and face > recognition would have to feed into it. > >> Le 06/10/2019 à 14:14, Germano Massullo a écrit : >>> Il giorno dom 6 ott 2019 alle ore 13:32 Moritz Mœller >>> ha scritto: Define 'most focused'. I give you an example to understand this request better. [...] >>> Yes you are right. but in your case, the couple is the main thing that >>> is moving in the picture. For my use case imagine I am taking photos >>> to people that are giving a talk. Some photos of the burst may be >>> blurred because I moved the camera while shooting, instead some other >>> shoots of the same burst could have less blur effect beause my hands >>> were not moving during its exposure time so the photo will have less >>> blur effect. >>> It would be great if an algoritm could detect the best shots ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
On Sun, 6 Oct 2019 15:02:39 +0200, =?UTF-8?Q?Aur=c3=a9lien_Pierre?= wrote: > That can be easily done by computing the L2 norm of the laplacian of the > pictures, or the L2 norm of the first level of wavelets decomposition > (which is used in the focus preview), and taking the maximum. > > As usual, it will be more work to wire the UI to the functionality than > writing the core image processing. Consider the case where the AF locks onto the background. This will likely result in a very large fraction of the image being in focus, but this will be exactly the wrong photo to select. Perhaps center-weighting, luminosity-weighting (if an assumption is made that the desired subject is usually brighter than the background, but not extremely light), skin tone recognition (with all of the attendant problems of what constitutes "skin tone"), and face recognition would have to feed into it. > Le 06/10/2019 à 14:14, Germano Massullo a écrit : >> Il giorno dom 6 ott 2019 alle ore 13:32 Moritz Mœller >> ha scritto: >>> Define 'most focused'. >>> I give you an example to understand this request better. [...] >> >> Yes you are right. but in your case, the couple is the main thing that >> is moving in the picture. For my use case imagine I am taking photos >> to people that are giving a talk. Some photos of the burst may be >> blurred because I moved the camera while shooting, instead some other >> shoots of the same burst could have less blur effect beause my hands >> were not moving during its exposure time so the photo will have less >> blur effect. >> It would be great if an algoritm could detect the best shots -- Robert Krawitz *** MIT Engineers A Proud Tradition http://mitathletics.com *** Member of the League for Programming Freedom -- http://ProgFree.org Project lead for Gutenprint --http://gimp-print.sourceforge.net "Linux doesn't dictate how I work, I dictate how Linux works." --Eric Crampton ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
That can be easily done by computing the L2 norm of the laplacian of the pictures, or the L2 norm of the first level of wavelets decomposition (which is used in the focus preview), and taking the maximum. As usual, it will be more work to wire the UI to the functionality than writing the core image processing. Le 06/10/2019 à 14:14, Germano Massullo a écrit : > Il giorno dom 6 ott 2019 alle ore 13:32 Moritz Mœller > ha scritto: >> Define 'most focused'. >> I give you an example to understand this request better. [...] > > Yes you are right. but in your case, the couple is the main thing that > is moving in the picture. For my use case imagine I am taking photos > to people that are giving a talk. Some photos of the burst may be > blurred because I moved the camera while shooting, instead some other > shoots of the same burst could have less blur effect beause my hands > were not moving during its exposure time so the photo will have less > blur effect. > It would be great if an algoritm could detect the best shots > ___ > darktable developer mailing list > to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org > ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
Il giorno dom 6 ott 2019 alle ore 13:32 Moritz Mœller ha scritto: > > Define 'most focused'. > I give you an example to understand this request better. [...] Yes you are right. but in your case, the couple is the main thing that is moving in the picture. For my use case imagine I am taking photos to people that are giving a talk. Some photos of the burst may be blurred because I moved the camera while shooting, instead some other shoots of the same burst could have less blur effect beause my hands were not moving during its exposure time so the photo will have less blur effect. It would be great if an algoritm could detect the best shots ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
Re: [darktable-dev] Automatically select the most focused photo in a burst of photos
Define 'most focused'. I give you an example to understand this request better. I shoot (tango) dancers in low light. A lot of shots are shoulders and heads only. Because I shoot in low light I use a very fast, manual lens at full aperture (f/0.95). In a burst sequence of seven shots that my Sony produces one will have the closest eye of the couple in focus. This only means a few pixels of the focus plane intersecting the face somehow. That's usually the best picture. The rest are for the bin. However, an adjacent picture that is completely useless may have the chest in focus which covers many more pixels than the face. How should darktable decide which is the 'most focused'? I think the only way to solve this well is with ML and possibly user input about the intent. No magic bullet exists for this type of problem. .mm ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org
[darktable-dev] Automatically select the most focused photo in a burst of photos
When you shoot a burst of photos, then in darktable you will select the one that is more focused (less blurred). It would be great if darktable would be able to select automatically the less blurred photo Related ticket: https://redmine.darktable.org/issues/12712 ___ darktable developer mailing list to unsubscribe send a mail to darktable-dev+unsubscr...@lists.darktable.org