We heard from many people that smartphones are still too complicated for many users, and even the simplest apps are too complicated for many users. I have talked to assistive tech trainers and rehab centers that even SeeingAI is too hard to use for their clients. It sounded like the current apps are still too complicated for maybe more than 75% of this community.
I think it is possible to develop a text reader app, that has all the functionalities but:
1) Has a much simpler user interface
2) Uses AI to predict the user's intention (such as between reading handwriting, a document or a short mail) and switch to that function automatically
3) Uses AI to fix angle, light, distance problems as much as possible to compensate for users who struggle to use the mobile camera perfectly
I feel like the audience in this forum is significantly more tech savvy than an average person with visual impairment, so maybe people here do not struggle with the problems we are trying to address. However, I was wondering if you still struggle with using the text reading apps or if you know anyone who struggles with them.
I would love to hear about your experiences here. We are also offering $10 gift cards for anyone who is willing to have a 30 minute phone interview about their pain points.
If you think I am misunderstanding something, I would love to hear about it too. Feel free to be very frank. We would rather focus on solutions that can make a difference than solutions that focus on fictional problems.
Comments
Reading printed
I don't find smart phones being complicated for day-to-day use. However on the specific issue of reading print or handwritten text using AI apps I agree that most of the time I fail to adjust the camera focus and hence don't get desired results however efficient the AI or OCR of such apps.
So hope to see this developing further!
Tripolice, thanks so much for
Tripolice, thanks so much for sharing your experience.
As far as I know, all of the text reading apps rely on you to take a good picture and then use AI or OCR to read that picture. They take a standard picture for whatever the condition is. Almost all of these AI systems were trained to read a regular picture in a reliable and robust way, but not trained to help the user take good pictures.
I think if AI was used to help that, then the rate of success could be much higher than the current rate.
I would welcome such an app.
I am a PhD researcher at the university of Edinburgh. I have to read many many books which have got Yellow Pages or basically in a tattered condition. Obviously, the problem I face most of the times, is to put my camera angle right. Many times I have to rescan a single page again and again and again.
Secondly at times I have difficulty reading diagrams or tables printed on the page. I would really really be grateful if such an app could be developed which would address these kinds of issues.
I would be also willing to take a telephone call from you, if you would like, to talk about my experiences further.
Voice Dream scanner simplest I've found
You can't get any simpler than this app for the price, and I speak as an AT trainer.
I second on Voice Dream Scanner and Seeing AI
Hi
Voice Dream Scanner and Seeing AI are truly excellent for reading with OCR in my opinion. I haven't done much processing of handwriting though. Of course I always welcome inovation, so if you have a great idea or can improve upon existing solutions, go for it! With regards to user interfaces of these apps in general, I don't think they are complicated at all, though I can imagine if people who get into the game at a later stage might think so. More than 75% strikes me as quite a high number though, but perhaps I'm just experiencing a different segment.
Best
Malthe
Well
As said in previous posts, the apps themselves are not the problem. I mean, give them the right picture and they'll give you the right feedback. However, an app that will guide us on how to Take that right picture can only be a super plus in my opinion. It can only get better, right?
Comments
I don't think the apps are complicated to use, but taking the picture makes it seem really hard. Even with guided access, it's not always easy to get the perfect picture, or maybe not the perfect picture, but it's not always easy to get to a point where you can get fairly accurate results
Thanks for the comments
Thanks so much for your comments. I would love to have a quick phone conversation with some of you. We are happy to give 10$ gift card for a 30 min conversation. Please email me at emre@mediate.tech if you are interested.
using tones?
Has anyone tried to use pitch tones to help adjust the camera angle?
Basically what you would use is a tone as a target sound, and another tone to help guide you in the right direction.
You would then know if your camera is lined up properly when all the tones sound the same.
I hope my explanation makes sense.
Tones, for me at least would be far quicker than speech guiding you every couple of seconds.
If only Apple used tones instead of this tilt left, tild right business, then again, they could keep both and let us choose which one we prefer.
AI to the resque?
I cannot read printed text myself, but maybe AI is the answer. Recently I read a lot of articles about apps that use this technology to improve the quality of pictures: the user just move the camera around and the software automatically takes a number of photografies, finally they are combined to crate the best image possible. The same principle could be used to take the best picture possible of a printed page, making it easier recongnizable by an ocr sistem, or as readable as it could be...
Just an idea, a crazy one perhaps...