How do I teach VoiceOver and iOS on iPhone 15 Pro to recognise people in photos? do I somehow name one and then next time it sees a person it will name them?
Not sure how accessible are the steps but you can try it and if it works let me know so I can do so. There are 2 ways. see below.
Find and name people and pets in Photos on iPhone
The Photos app sorts photos and videos of people, dogs, and cats in the People & Pets collection (or the People collection if no pets appear). You can name a person, dog, or cat directly from a photo or video in your photo library, or in People & Pets.
Name a person or pet in a photo or video
When you name a person or pet in a photo or video, theyâre automatically added to People & Pets and named in other photos and videos in your library.
Go to the Photos app on your iPhone.
Open a photo or video, then swipe up to see the details.
People or pets in the photo or video appear in the bottom-left corner. A question mark appears next to those you havenât named.
Tap the person or pet with a question mark next to their picture, then tap Name This Person or Name This Pet.
The top half of the iPhone screen shows a photo open in the Photos app. In the bottom-left corner of the photo are question marks next to the people and pets that appear in the photo. A question mark is selected and a menu of options includes, from top to bottom, Name This Person, All Photos, Make Key Photo, Edit Name and Photos, and Feature This Person Less.
Enter their name, then tap Done.
Name a person or pet in the People & Pets collection
Go to the Photos app on your iPhone.
Below the photo library, tap People & Pets, then tap the person or pet you want to name.
The People & Pets collection in the Photos app. Groups appear at the top, and people and pets are listed below.
Tap Name This Person or Name This Pet, then enter their name.
I could be wrong, from what I remember, your device generally recognizes people and pets from your pictures, and then you can name them. I don't think there is a way to do this intentionally though. For example, I might have pictures of people in my phone, but for some reason, these are not recognized and sorted in the people Tab on my device. When my iPhone doesn't recognize them, or give me the option to give this individual a name in my phone, then it's not going to allow me to manually input this data. It makes no sense. The problem here is, you would have to know who is in the picture, and, I'm guessing that would be hard to do if you're totally blind. I don't know if this has changed since the update of iOS 18, but if your device doesn't initially recognize the people in your pictures, then it's hard for you to give them a name if they don't show up in the people or pets album. Another major problem with the system is that, there is no way for you to, for example, get a single picture of the individual you want to add to your device, and then add them to the people album on the phone. I think there was an application, I don't remember if it was seeing AI, that would allow you to train the application directly to recognize the people in your pictures. The problem is, you have to be able to take three consecutive photos of the person. This means that you would have to get help from the person in order to take the pictures, at least, initially. When you're trying to set this up, they even tell you that on the screen. Yes, I understand that this is a privacy issue, according to them. What if it's a celebrity or a public figure? You can't take three single pictures of them from the Internet, and use it to train your device regarding the information or the data in your phone, for when you download pictures of them onto your device. I know that, as a totally blind person, no matter how much technology is available, there are going to be limitations. A sighted person can take as many pictures of somebody who they want to, with or without their knowledge, but we can't do that. If I have a picture, or multiple single pictures, of a family member, I cannot just import that into my device, and train them to recognize this individual. No, at least with seeing AI, you definitely had to take the pictures right then and there. What if the person died or something, but you still have pictures of them? Shouldn't you be able to train your device to recognize them, so that you can go through the pictures of them for the sake of memories? I don't get it. Well, no, that's not true. I get it, but it's still incredibly frustrating. Do you honestly think I'm going to have pictures of random strangers on my phone that I'm trying to have recognized? I don't know. I understand that privacy is an issue, but, I hate the limitations of blindness sometimes, well, all the times.
I understand the concept of giving each individual in your phone a name. What I'm asking is, I know that when the people come up, you're usually shown a picture of their faces, and then you can decide, OK, face one is Aunt Jennifer, phase 2 is uncle John. What if you have somebody else in your device, Cousin Samuel, but your device does not show that person in the people section, like their face does not come up. How do you add that person to your device to be recognized in the future, if the device does not initially give you the option to do so? That's what I'm trying to figure out. That's the problem here. Plus, if you're totally blind, initially when you're setting this up, you're most likely going to need someone to go through the pictures with you, identify the faces that you have recognized on the device, and then put the name of them in, as well as combine faces that are the same person, that your device probably thought was a different person
I know it's not ideal but it might have to require a more logical approach, IE, knowing the content and people in the picture. Start with a picture of yourself, name > Include a picture of you and a relative, you should already be named and then label them > etc. By this process you can label those who occur regularly in your images.
It's certainly not perfect, but I do imagine it could be a useful tool once set up. There is, of course, the option that is the unspoken one in our community, ask for someone to help! It's fun going through photographs with people, I find anyway.
Unfortunately, I can't imagine an easier way this could work without some really awkward security issues. Maybe Apple could allow people to share you their image ID, so, for example, my friend Hazel sends me a picture of herself which is already labelled with her being in the image, I can then add her to my people.
Just to check, if a photograph is shared, are identities included in the meta data? I don't think it is, but I could be wrong... Often am.
I did it in the past.
I don't remember how, but you can add a discription to a photo.
Like this VoiceOver recognize people in photo.
You can lable it but it doesn't sa( the the operating system is updating.
You can look for it if you need it in the future.
,taking pictures directly in not necessary. digital photo can be done, too, all you need is two devices. use one device to display the photo, then take pictures from the other device using seeing ai.
when the photo have multiple people on it, you need to focus on just one person, so you can't see the photo in the first place, you'll have problem as you need to know where is the person you want and then focus on it.
Neither of these two last adress the OP problem which is differenciating between subjects.
Yes, you can add a description with mark up but on IOS 18, and before, it is possible for the photos app to be trained to reckognise people meaning new photographs that come into the app are automatically sorted and identified based on subjects.
I believe what Winter Roses is asking for, is a similar feature that is currently in Seeing AI, to be added here. In Seeing AI there is a Face Recognition feature that allows you to take a selfy of someone, and add a name to it. Then when you activate the recognition feature later, you may get something like:
> A person to the right of the frame. . .
> a person center of the frame. . .
> John is to the left of the frame. . .
I believe this functionality may also carry over to Seeing AI's scene recognition, where you may get something like:
> A person sitting in a bean bag chair. . .
> A woman lounging on a couch. . .
> John is sitting in a rocking chair. . .
Hi,
This can indeed be done, but if there are multiple people in the photo, you may need sighted assistance to be sure youâre naming the right one, and the results are not always as hoped..
Open Photos and open the photo with the person you wish to name.
Select the Info button at the bottom, next to Share. Or select Show Details in the rotor.
Swipe all the way past all the technical camera related information to find an item labelled âUnnamed personâ.
Double tap and hold to open a pop up menu.
Select âName this personâ.
Type their name and select Done.
And Bobâs your uncle.
In theory, if a person already has a name, even recognised from a different photo, this will appear instead of it saying âUnnamed personâ, as you would want.
And when browsing photos, VoiceOver will say something like âmaybe Davidâ. Again though, I donât have complete success with this. Biut perhaps it needs time to process or something.
Hope that helps and happy tinkering,
Dave
I'm glad this topic came up, as I knew this was now possible, but hadn't gotten around to trying it yet.
I gave it a try yesterday, and found it rather difficult, even as a low vision user. I could get to the point of finding the unknown person buttons, but had no idea which one was which. Sometimes it seemed like there was a glowy outline around the person when I tapped one of the unknown buttons, but it was inconsistent.
So, here's my idea to make this more helpful for both blind and low vision users...
I think that when a VoiceOver user puts focus on one of the unknown person buttons, it should announce the rough position of where they are in the photo, like person on the far left, or person in the back row second from the far right. This would still be rather difficult to name people in a photo with loads of people, but for things like a family photo with 3 to 5 people, it wouldn't be too bad.
Then the user could first open a photo, recognize or describe it with BE My Eyes or something, and learn that there's a tall man in the middle,a younger female child on the right, and a smiling woman on the left. With that knowledge, and VoiceOver telling the position of which person is in focus, it would be easy to name your aunt, uncle, and cousin...
I'm not suggesting that we replace entirely what we have now, but, yeah. That's the general idea. I was thinking that, if privacy is a concern, well, give us the option to be able to have somebody take three pictures of themselves, where their face is clearly visible, in our phone. Then, using something like artificial intelligence, for example, your iPhone could store their face. Give us the option to label their face in the camera roll in real time. This would help identify the faces easier. Rather than going through a family picture, Trying to choose the faces from the list, I'm suggesting that we have something a little bit more intentional. There's nothing wrong with doing it that way, if you're able to, and if you have the help to identify who is in the pictures initially. I'm saying that, it would be nice if someone could send you three pictures of themselves, single pictures would work best, where their face is visible, and then you can train your phone to identify them, Name them, and then the process would be done. As I mentioned before, I understand that, for privacy reasons, they might not want you using digital pictures. Anybody could go on the Internet, get photos, and train their device to recognize this person. Again, this would really be a barrier to us, as blind people, not for sighted people. Ideally, in the interest of privacy, the individual would have to take the three pictures of themselves on your phone directly. They wouldn't be able to send you them digitally, although this would be great if they allowed this. I mean, most people have an idea of who they have in their phone, or, at the very least, who they would like to be identified. This means that, your cousin Jennifer, could take three pictures of herself, pictures with only her, either take them directly on your device, or send them to you, and then you could train your phone to recognize her. Then, you could do the same thing for Uncle Samuel, and so on. This would eliminate the need for you to go through your device, find out who has already been identified by your device, and then give them a name. The process would be easier because, it would be fairly easy to obtain a picture, or a few pictures, with the people that you want to be identified. If you don't have any single pictures of them, then, you would have to get someone to help you crop a picture, to include only them, so that your device can identify them easier. Just a suggestion. Let me know what you think, if this makes sense.
The thing is, your device doesn't always recognize every single face in your pictures, especially if they are not clear. I don't think that the names are going to transfer to different applications. It would be nice if Be My Eyes implemented a feature similar to what seeing AI already has. Again, for privacy reasons, I don't think the third-party applications are going to be able to access the name of the persons on your device, because this would make it easier on us as well. I like the idea of adding a description to the photo afterwards, in conjunction with identifying who is in the picture.
Comments
Naming people or pets
Not sure how accessible are the steps but you can try it and if it works let me know so I can do so. There are 2 ways. see below.
Find and name people and pets in Photos on iPhone
The Photos app sorts photos and videos of people, dogs, and cats in the People & Pets collection (or the People collection if no pets appear). You can name a person, dog, or cat directly from a photo or video in your photo library, or in People & Pets.
Name a person or pet in a photo or video
When you name a person or pet in a photo or video, theyâre automatically added to People & Pets and named in other photos and videos in your library.
Go to the Photos app on your iPhone.
Open a photo or video, then swipe up to see the details.
People or pets in the photo or video appear in the bottom-left corner. A question mark appears next to those you havenât named.
Tap the person or pet with a question mark next to their picture, then tap Name This Person or Name This Pet.
The top half of the iPhone screen shows a photo open in the Photos app. In the bottom-left corner of the photo are question marks next to the people and pets that appear in the photo. A question mark is selected and a menu of options includes, from top to bottom, Name This Person, All Photos, Make Key Photo, Edit Name and Photos, and Feature This Person Less.
Enter their name, then tap Done.
Name a person or pet in the People & Pets collection
Go to the Photos app on your iPhone.
Below the photo library, tap People & Pets, then tap the person or pet you want to name.
The People & Pets collection in the Photos app. Groups appear at the top, and people and pets are listed below.
Tap Name This Person or Name This Pet, then enter their name.
Tap Done.
Not sure if this is possible by training
I could be wrong, from what I remember, your device generally recognizes people and pets from your pictures, and then you can name them. I don't think there is a way to do this intentionally though. For example, I might have pictures of people in my phone, but for some reason, these are not recognized and sorted in the people Tab on my device. When my iPhone doesn't recognize them, or give me the option to give this individual a name in my phone, then it's not going to allow me to manually input this data. It makes no sense. The problem here is, you would have to know who is in the picture, and, I'm guessing that would be hard to do if you're totally blind. I don't know if this has changed since the update of iOS 18, but if your device doesn't initially recognize the people in your pictures, then it's hard for you to give them a name if they don't show up in the people or pets album. Another major problem with the system is that, there is no way for you to, for example, get a single picture of the individual you want to add to your device, and then add them to the people album on the phone. I think there was an application, I don't remember if it was seeing AI, that would allow you to train the application directly to recognize the people in your pictures. The problem is, you have to be able to take three consecutive photos of the person. This means that you would have to get help from the person in order to take the pictures, at least, initially. When you're trying to set this up, they even tell you that on the screen. Yes, I understand that this is a privacy issue, according to them. What if it's a celebrity or a public figure? You can't take three single pictures of them from the Internet, and use it to train your device regarding the information or the data in your phone, for when you download pictures of them onto your device. I know that, as a totally blind person, no matter how much technology is available, there are going to be limitations. A sighted person can take as many pictures of somebody who they want to, with or without their knowledge, but we can't do that. If I have a picture, or multiple single pictures, of a family member, I cannot just import that into my device, and train them to recognize this individual. No, at least with seeing AI, you definitely had to take the pictures right then and there. What if the person died or something, but you still have pictures of them? Shouldn't you be able to train your device to recognize them, so that you can go through the pictures of them for the sake of memories? I don't get it. Well, no, that's not true. I get it, but it's still incredibly frustrating. Do you honestly think I'm going to have pictures of random strangers on my phone that I'm trying to have recognized? I don't know. I understand that privacy is an issue, but, I hate the limitations of blindness sometimes, well, all the times.
Winter Roses
You need to name them.
Question
I understand the concept of giving each individual in your phone a name. What I'm asking is, I know that when the people come up, you're usually shown a picture of their faces, and then you can decide, OK, face one is Aunt Jennifer, phase 2 is uncle John. What if you have somebody else in your device, Cousin Samuel, but your device does not show that person in the people section, like their face does not come up. How do you add that person to your device to be recognized in the future, if the device does not initially give you the option to do so? That's what I'm trying to figure out. That's the problem here. Plus, if you're totally blind, initially when you're setting this up, you're most likely going to need someone to go through the pictures with you, identify the faces that you have recognized on the device, and then put the name of them in, as well as combine faces that are the same person, that your device probably thought was a different person
I know it's not ideal but itâŠ
I know it's not ideal but it might have to require a more logical approach, IE, knowing the content and people in the picture. Start with a picture of yourself, name > Include a picture of you and a relative, you should already be named and then label them > etc. By this process you can label those who occur regularly in your images.
It's certainly not perfect, but I do imagine it could be a useful tool once set up. There is, of course, the option that is the unspoken one in our community, ask for someone to help! It's fun going through photographs with people, I find anyway.
Unfortunately, I can't imagine an easier way this could work without some really awkward security issues. Maybe Apple could allow people to share you their image ID, so, for example, my friend Hazel sends me a picture of herself which is already labelled with her being in the image, I can then add her to my people.
Just to check, if a photograph is shared, are identities included in the meta data? I don't think it is, but I could be wrong... Often am.
You can add a discription to a photo.
I did it in the past.
I don't remember how, but you can add a discription to a photo.
Like this VoiceOver recognize people in photo.
You can lable it but it doesn't sa( the the operating system is updating.
You can look for it if you need it in the future.
photos can be used to train person recognition in seeing ai,
,taking pictures directly in not necessary. digital photo can be done, too, all you need is two devices. use one device to display the photo, then take pictures from the other device using seeing ai.
when the photo have multiple people on it, you need to focus on just one person, so you can't see the photo in the first place, you'll have problem as you need to know where is the person you want and then focus on it.
Neither of these two lastâŠ
Neither of these two last adress the OP problem which is differenciating between subjects.
Yes, you can add a description with mark up but on IOS 18, and before, it is possible for the photos app to be trained to reckognise people meaning new photographs that come into the app are automatically sorted and identified based on subjects.
Seeing AI feature...
I believe what Winter Roses is asking for, is a similar feature that is currently in Seeing AI, to be added here. In Seeing AI there is a Face Recognition feature that allows you to take a selfy of someone, and add a name to it. Then when you activate the recognition feature later, you may get something like:
> A person to the right of the frame. . .
> a person center of the frame. . .
> John is to the left of the frame. . .
I believe this functionality may also carry over to Seeing AI's scene recognition, where you may get something like:
> A person sitting in a bean bag chair. . .
> A woman lounging on a couch. . .
> John is sitting in a rocking chair. . .
Hope this makes sense. đ
How to name people
Hi,
This can indeed be done, but if there are multiple people in the photo, you may need sighted assistance to be sure youâre naming the right one, and the results are not always as hoped..
Open Photos and open the photo with the person you wish to name.
Select the Info button at the bottom, next to Share. Or select Show Details in the rotor.
Swipe all the way past all the technical camera related information to find an item labelled âUnnamed personâ.
Double tap and hold to open a pop up menu.
Select âName this personâ.
Type their name and select Done.
And Bobâs your uncle.
In theory, if a person already has a name, even recognised from a different photo, this will appear instead of it saying âUnnamed personâ, as you would want.
And when browsing photos, VoiceOver will say something like âmaybe Davidâ. Again though, I donât have complete success with this. Biut perhaps it needs time to process or something.
Hope that helps and happy tinkering,
Dave
Naming Photos Idea
I'm glad this topic came up, as I knew this was now possible, but hadn't gotten around to trying it yet.
I gave it a try yesterday, and found it rather difficult, even as a low vision user. I could get to the point of finding the unknown person buttons, but had no idea which one was which. Sometimes it seemed like there was a glowy outline around the person when I tapped one of the unknown buttons, but it was inconsistent.
So, here's my idea to make this more helpful for both blind and low vision users...
I think that when a VoiceOver user puts focus on one of the unknown person buttons, it should announce the rough position of where they are in the photo, like person on the far left, or person in the back row second from the far right. This would still be rather difficult to name people in a photo with loads of people, but for things like a family photo with 3 to 5 people, it wouldn't be too bad.
Then the user could first open a photo, recognize or describe it with BE My Eyes or something, and learn that there's a tall man in the middle,a younger female child on the right, and a smiling woman on the left. With that knowledge, and VoiceOver telling the position of which person is in focus, it would be easy to name your aunt, uncle, and cousin...
Just a thought...
Naming persons
I'm not suggesting that we replace entirely what we have now, but, yeah. That's the general idea. I was thinking that, if privacy is a concern, well, give us the option to be able to have somebody take three pictures of themselves, where their face is clearly visible, in our phone. Then, using something like artificial intelligence, for example, your iPhone could store their face. Give us the option to label their face in the camera roll in real time. This would help identify the faces easier. Rather than going through a family picture, Trying to choose the faces from the list, I'm suggesting that we have something a little bit more intentional. There's nothing wrong with doing it that way, if you're able to, and if you have the help to identify who is in the pictures initially. I'm saying that, it would be nice if someone could send you three pictures of themselves, single pictures would work best, where their face is visible, and then you can train your phone to identify them, Name them, and then the process would be done. As I mentioned before, I understand that, for privacy reasons, they might not want you using digital pictures. Anybody could go on the Internet, get photos, and train their device to recognize this person. Again, this would really be a barrier to us, as blind people, not for sighted people. Ideally, in the interest of privacy, the individual would have to take the three pictures of themselves on your phone directly. They wouldn't be able to send you them digitally, although this would be great if they allowed this. I mean, most people have an idea of who they have in their phone, or, at the very least, who they would like to be identified. This means that, your cousin Jennifer, could take three pictures of herself, pictures with only her, either take them directly on your device, or send them to you, and then you could train your phone to recognize her. Then, you could do the same thing for Uncle Samuel, and so on. This would eliminate the need for you to go through your device, find out who has already been identified by your device, and then give them a name. The process would be easier because, it would be fairly easy to obtain a picture, or a few pictures, with the people that you want to be identified. If you don't have any single pictures of them, then, you would have to get someone to help you crop a picture, to include only them, so that your device can identify them easier. Just a suggestion. Let me know what you think, if this makes sense.
The thing is, your device doesn't always recognize every single face in your pictures, especially if they are not clear. I don't think that the names are going to transfer to different applications. It would be nice if Be My Eyes implemented a feature similar to what seeing AI already has. Again, for privacy reasons, I don't think the third-party applications are going to be able to access the name of the persons on your device, because this would make it easier on us as well. I like the idea of adding a description to the photo afterwards, in conjunction with identifying who is in the picture.