Primary navigation

Unlocking the Power of Reading with Microsoft's Seeing AI: A Guide for Partially Sighted Users

Shaun Preece:
As a great philosopher once said, "We just got a letter. We just got a letter. We just got a letter. I wonder who it's from." Good question because I have no idea at all, but there is a way around that.

Announcer:
Shaun of the Shed, an AMI original podcast.

Shaun Preece:
How do you do, and welcome to another episode of Shaun of the Shed. I am that Shaun and this is my shed, and this is the tech show where I try and show you just how useful technology can be because it is useful, but more than that, it's life-changing, or at least it can be. No, but honestly, it can. Sarah, it can be. They're not toys. They're not gadgets. I'm not in here just playing with toys, Sarah. This is my job. This is what I do. It's what-

Announcer:
We'll be right back after Shaun's nappy time.

Shaun Preece:
Things got personal there. Post, mail, snail mail, call it whatever you like. I call it the devil's lavatory paper because I hate the post. I mean, it's just another thing you got to deal with every day, and if you can't see printed text, it's even worse. Of course, it is. Maybe you're blind or visually impaired or maybe you've got some other disability, dyslexia, maybe, meaning that you find it really difficult to read text. What do you do then? Don't worry, until we actually abolish the post, which I'm all for by the way. It needs to go the same way as the fax or leg warmers, I don't know, but until it's gone, there is a way we can manage the post. And of course, it comes down to technology. Oh, gorgeous, gorgeous technology.
In fact, it's as easy as pulling your smartphone out of your pocket. All you need is the right software. So that's what I'm going to do today. I'm going to show you some software for both an iPhone and an Android phone that will enable you to read your printed text, or in this case a letter. You simply point the camera at what you want to be read and your smartphone will read it out to you in a cool, natural-sounding voice. It's really good. Now, it's not just your smartphone that can do this. Any computer, as we all know, a smartphone is basically a small computer in your pocket, but your laptop or your desktop computer at home can also do this, and as long as it has a built-in camera. This whole process is called OCR, which stands for Optical Character Recognition. Sounds jargony, right? Yes, it is and it is incredibly tedious, actually. We don't have to go into that, so forget about what I just said.
Now all you need is the right software to do this. So I'm going to download for my iPhone an app called Seeing AI. Now, of course, it's not the only app that can do this. There's plenty of others out there, Envision, Supersense, Voice Dream Scanner. There's lots of different options, but Seeing AI is my favorite, and it's not just because it's free. No. No, no. Well, yes, okay, it has a big deal to do with it. It's free. I love that. Seeing AI can do lots of other things as well, not just reading printed text, but we'll get into that another time. So let's download Seeing AI. Grab your iPhone, and you need to find the app store. Once you do, tap to open it. Now in the bottom right corner, you'll find the search tab. Tap on that to select it.
Now in the top of the screen you'll find a search field and type in there Seeing AI. That's S-E-E-I-N-G, space, A-I, which by the way stands for artificial intelligence. Hit enter and the first result should be Seeing AI from Microsoft. Now, I've already got this installed, so my button says open, yours will say download or get it or whatever it says. Either way, get it and then we'll open it in just a second and test it out. Okay, so I have in my hand a piece of paper. Well, okay, it's a letter, but obviously it's not one of my real letters. No one wants to read that. It'd just be a bill and then we'd all get depressed and then I'd start crying. We just don't need that. So this is a letter that I just made up and printed out myself, but it doesn't matter because the process will still be the same.
So what I'm going to do is just flatten out this letter on a flat surface, my desktop just here, and then let's open up Seeing AI go through the process. Just like with any other app, when you run it for the first time, there is a certain amount, I'm going to use the phrase fluff and bumf that you're going to have to go through, such as agreeing to the terms of use, allowing access to the camera. Also, there's some tips and help and advice as well, but once you go through that after the first time, you don't have to do it again. Once you actually get to the main screen of Seeing AI, it's actually really simple. There's not much here. Let's go through it. So in the top left corner, the first item is menu. Now you can go in here to change various settings, but there's nothing too important in here. If we swipe onto the next item, it is quick help.
Every time you select a function, and we'll get to that in just a minute, of Seeing AI, it will give you a quick, brief description of what that function does and some helpful advice. Quick help, there's clues in the name, makes sense. The next item we've got here is the recognizing language. So if you're reading text that is in a different language other than English, then you can select that here. Now the thing that's really important is this final item on the screen and it is the channel selection. Channel, I know sounds strange, but that's what they call each feature or function of Seeing AI.
And as I said earlier, there's various things that Seeing AI can do, not just reading printed text, but obviously that's what we're going to focus on today. So once you hear the channel, you can swipe up or down to choose which channel you want to use. Now the first one, at least on here, is something called short text. Now we will come back to that later on. Let's swipe up with one finger and the next item in the channel is document. Now, obviously that's the one I want to use. Now, once we've had that selected, you don't have to do anything else other than here it's a document. We're going to place the phone on top of the document, as I mentioned earlier, and I'm going to lift it up and see how well the document is positioned in the camera.

Device:
Left edge not visible.

Shaun Preece:
Now as you can hear, as I'm lifting it up, it gives me guides, "Left edge, not visible," so I know I need to move the phone a bit to the left, and as I said earlier as well, look, this can be a little bit annoying and does take a while to get the hang of this, but honestly you will get it.

Device:
Hold steady. Processing.

Shaun Preece:
There. Now, as you heard, as soon as it is in frame and Seeing AI can recognize some text, it will automatically take a picture, and once it does that, it will take us onto the text screen where it will read what text its recognized and what the letter says. Now there's a few... well, actually there's two different ways we can get it to read this text. As I have the screen reader on on my iPhone, I can simply swipe through until I get to the text. Let's listen to it.

Device:
Mr. Shaun Preece. One The Shed. Bottom of the Garden. Shed on Sea. SH3 D13. Dear Sir, Thank you for your recent correspondence requesting that to your property should be recognized and officially listed as a building of national significance. In response to some of your statements, I would like to draw your attention to the following points. One, while it is true that the birthplaces of John Lennon, William Shakespeare, Charles Dickens, and Benny Hill are all officially recognized as locations of national significance, sadly, you're claim that you have a YouTube channel and therefore are, to use your words, "Basically on the same level as that there Stephen Fry," somewhat questionable. Two, usually when a plaque declaring a building of the birthplace or residents of a national figure is put in place, it is attached to a actual bricks and mortar building. As of yet, no such plaque has ever been awarded to a garden shed. Three, you live in a garden shed. Stop it. Play button. Share button.

Shaun Preece:
So as you can hear, that was pretty much perfect. You could understand every word that was recognized. That is really good, and to be honest, I find that it is really accurate pretty much all of the time. Now, if I keep swiping on, you'll find at the bottom there is a play button. If you don't use a screen reader with your iPhone, you can press or tap the play button and it will read the text it's recognized to you using an inbuilt voice. So there it is. As easy as that, as brilliant as that. I mean, it just makes it so easy. I absolutely love it. Also, at the bottom here you'll find a share function, so if you want to email this to someone or whatever you want to do, print it out, maybe, you can do that using that feature.
Now, let's go back here. The top left corner, there's a back button which will take us back to the main screen if we wanted to scan another document, but let's take a look at the short text. If I swipe down on the channel selection area again, it will take us back to the short text function.

Device:
Short text.

Shaun Preece:
This is a really cool function, but it is slightly different to the document and you'll see why in just a second when I demo it. But basically what it does, and the name sort of gives it away, it is for reading shorter amounts of text. And I sort of use this as almost a way to skim through text and documents easier or faster than using the document way because if you use the document channel, as you heard, it will take a picture of the document, whereas if you use short text, as soon as it sees anything, any text that is, in the viewfinder of the camera, it will just read it aloud straight away. It doesn't wait to take a picture, it just tries to read what it can see in the camera.
Now, obviously it's a lot faster, but it can be less accurate, but it's fine. You can use it, as I say, as a way to skim through documents. If I skim through a letter and I hear it say, I don't know, cabbage, cabbage, gas bill, 500, cabbage, carrot, I know, "Okay, so it might not have got the other words right, cabbage, cabbage, but I know that's my gas bill and it's time to cry." So this is short text. Let's have a quick demo of that now.

Device:
Short text.

Shaun Preece:
Okay, so now I've got the short text channel selected. If I just place the phone somewhere over the document, you'll hear it starts reading whatever text it picks up straight away.

Speaker 4:
Thank you for your recent correspondence requesting that your property should be recognized and officially list... the shed.

Shaun Preece:
There we go. How quick was that? No need to take a picture, no need for it to be totally all in the frame. It just reads whatever it can see. Fantastic. Now you will notice that if I just move the phone a little bit...

Speaker 4:
T, Mr. Shaun, the shed.

Shaun Preece:
... then it will restart the text recognition. Whatever it sees again, it will say, "Oh, there's some new text that I've recognized. Let me start reading that aloud." So you do need to keep your phone pretty still to read all of the text, but that's not really what it's for. As the name suggests, it's for short amounts of text. And again, it's really good just for skimming through things and excellent if you want to read the label of a can or a food package or anything really. I use this short text function quite a lot for reading messages on my computer. If I turn my computer on and there's some message, "Windows is updating," it's quite handy for things like that. Anyway, that's short text and that's document channel on Seeing AI. Fantastic. It can make a huge difference.
Now, if you're not an iPhone owner, what have you got for Android? Well let's take a look. Okay, you gorgeous Android people. Of course, you're not out of the game. You can do this as well. There are plenty of apps for the Android platform that can do the same thing. Now the one I'm going to use is called Google Lookout. It's a great app and very much in the same vein as Seeing AI. They're very similar. But as I said, there are many other options of software when it comes to this. Again, Envision, I think Supersense is on Android as well, but this is the one which I find the most useful. So first thing we need to do is download it. So grab your Android phone, whatever it may be, and open the Play Store. Once there, search for Google Lookout all separate words. You should find it in the first result and just tap on the install button to install it and that's it. Easy.
Now, just like we're Seeing AI, the first time you actually open and run the Google Lookout app, you will have to contend with the, remember what it's called, the fluff and the bumf. Yes, you will have to give it permission to use the camera and agree to the terms and conditions and the various other things, but once you get through all that, let's go through the main screen of Google Lookout. Now, many of the items on this screen, you're going to recognize as very similar, as I said, to the items on Seeing AI. So let's quickly run through them. The very first thing on the top of the screen is help. If we swipe on, the next thing is the recognition language, and again, swipe and this is my Google account. Let's swipe on again. Here we have the option to turn the camera off for privacy reasons or whatever until you're ready to use it. If we swipe on again, the next item is recents.
Now, this is really cool and I wish Seeing AI had this. So if you've scanned a few documents previously, they will be in here and you can go back and browse them at your leisure. Really cool. Now the next items on the screen are the things we're really interested in, so if we swipe on, we have explore. Now this is a different function to the text that we're looking for, and again, we may come back to this at later episodes, but let's swipe on for now. Okay, food labels, again, we'll come back to this. Swipe on, and here is the one that's selected by default, text. Now this is a direct comparison to the short text channel in Seeing AI. So you know what? Let's quickly try this one out. If I just hold the phone over the letter now...

Speaker 5:
This automatic pet feeder one zero [inaudible 00:15:50].

Shaun Preece:
There, it will try and recognize whatever it sees straight away in real-time. For now, let's leave that alone and swipe on to the document feature.

Speaker 5:
Documents.

Shaun Preece:
Now again, like the document feature on Seeing AI, if we just hold the phone over the letter, it will guide us to get the best picture.

Speaker 5:
Too close, move device away. Hold still.

Shaun Preece:
Okay, move it away.

Speaker 5:
Feeding setup. This automatic pet feeder supports 0.19 inch minus 0.47 inch.

Shaun Preece:
Now, we won't go through it all, but as you heard, it did a great job, just like Seeing AI, of recognizing all the text in that letter, and I don't actually think it got anything wrong whatsoever. Very much comparable to Seeing AI. Fantastic. Fantastic.
Okay, you sweet potatoes you, let's leave it there. Hopefully, I've shown you just how, well, useful that feature is, but also how fairly easy it is, fairly straightforward. It is something I'm... I mean I joked at the beginning about how I hate post and letters and everything, but you know what? It's something we do have deal with and if you can't read it for yourself, these apps are just invaluable. Also, as I said, don't forget if you just have a laptop or a desktop computer, you can do the same thing with those. You can get webcams, you can use, or even document cameras specialized for doing this sort of thing. Or even if you have a all-in-one printer which has a scanner built in, you can do the same thing with the right software. Anyway, I hope you found this useful. If you did, please like and subscribe, and if you want to get in touch with me and tell me what I should look at next, you can. Just email feedback@doubletaponair.com. Thank you so much for watching guys, and I will see you next time.