2019-03-14 17:43:06

Hi everyone!

I've made an NVDA add-on called "Image describer" that takes the current navigator object and then says what it contains by looking at it visually. Kind of like the Facebook auto alt text, but this works anywhere.

The shortcut to interpret an image is NVDA+Control+I.

The add on can be downloaded at http://www.screenreader.ai/

I'm a sighted machine learning researcher that recently have caught an eye on the field of web accessibilit

Websites are primarily designed to be interacted with visually which is hard to do if you have a visual impairment. Here I think there's big potential for AI to help and look at the web page visually and then give correct and helpful information to the visually impaired or blind user.

I'm looking towards making much more than this, please contact me or reply to this thread if you want to get help with something accessibility wise or have any kind of feedback.

2019-03-14 17:55:13

awesome! I will give this a try
there was something similar but I guess it really never took off

A bus station is where a bus stops. A train station is where a train stops. On my desk, I have a work station…

2019-03-14 18:00:24

wow, cool!
it would be more awesome if you removed iran from the blocklist, it gives me access denied.

2019-03-14 18:21:48

Sounds cool, downloaded it!

2019-03-14 18:26:51 (edited by Destranis 2019-03-14 18:27:16)

Sorry for doubleposting. Would be good if it didn't say only text, but the text itself, too...

2019-03-14 18:38:50

will this be on the NVDA-addons site?

PSN ID: AvidLitRPGer
Twitter: https://twitter.com/AvidLitRPGer
Facebook: https://www.facebook.com/AvidLitRPGer
leave me a message saying how you found me.

2019-03-14 18:42:08

It will in the future. Right now it is under review, as can be seen on the NVDA addons mailing list.

2019-03-14 18:49:17 (edited by kianoosh 2019-03-14 18:55:58)

Yeah. It doesn't give me access to the website just cause i'm from iran. I can get the add on someway, but If that's gonna use an api on your website I can't use it unless I always keep a vpn connection on.
Edit: After getting the add on I see that it is using a website that blocked iran since I get the error 403. This is an unfortunate thing.

---
Co-founder of Sonorous Arts.
Check out Sonorous Arts on github: https://github.com/sonorous-arts/
my Discord: kianoosh.shakeri2#2988

2019-03-14 19:11:09

Well now this is nearly useless because it doesn't tell more than text, screen and stuff like that. Even facebook tells more. But i hope it'll improve and i'll keep an eye on it. Keep up the good work.

2019-03-14 20:20:13

well is it safe? no viruses?

Yours kindly

2019-03-14 20:56:09

Of course it's safe. I understand your concern, but such things can often discourage developers. The review is related to other things, not only security.

2019-03-14 21:12:41

this would be useful for video editting.

be a hero and stop Coppa now!
https://docs.google.com/document/d/1Dkm … DkWZ8/edit
-id software, 1995

2019-03-14 22:34:11

UltraLeetJ wrote:

awesome! I will give this a try
there was something similar but I guess it really never took off

Cool, that's awesome! Do you know remember name of that add-on?

2019-03-14 22:38:13

AlirezaNosrati wrote:

wow, cool!
it would be more awesome if you removed iran from the blocklist, it gives me access denied.

Hi! I'm sorry that you can't use it but right now it's out of my control, most major cloud providers forces you to block Iran because of US sanctions. If you want, send me an email at [email protected] and I can work something out for you.

2019-03-14 22:41:21

MasterOfDeath wrote:

Sorry for doubleposting. Would be good if it didn't say only text, but the text itself, too...

Right now the add-on is aimed to describe natural scene photographs taken from the real world, the thought is that when encountering text you would use the OCR add-on. If there is need I could definitely also add OCR capability if it detects text in the image. Or have I missed something?

I'm sighted and don't use a screen reader on a daily basis like most here do, your feedback is extremely valuable.

2019-03-14 22:43:49

ammericandad2005 wrote:

this would be useful for video editting.

Cool, didn't think about that. Would love to know more about applying this for video editing, if you want and have the time, send me an email at [email protected], your feedback is very valuable for me since I have regular sight.

2019-03-14 23:49:59

oliver.edholm wrote:

Cool, that's awesome! Do you know remember name of that add-on?

I think it was called "describeIt"
there is info on that link. I think that one does OCR

A bus station is where a bus stops. A train station is where a train stops. On my desk, I have a work station…

2019-03-15 02:21:57 (edited by Destranis 2019-03-15 02:24:22)

Well, you could add that of course, and that would be really useful for many of us. OCR is sometimes not that accurate and if you could also implement that somehow, maybe it would improve, together with the image recognising feature. But as a starting, it's absolutely good, and if you keep developing it, we'll appreciate it. smile

2019-03-15 02:58:47

grate !. I can use in facebook!. especially when there are ! pictures  NVDA can tell what the picture is

2019-03-15 06:42:26

@14 i don't need the add-on since i can see, but there are many iranians who need it.

2019-03-15 08:54:27

oh very good addon!
if you are gonna continue developing, we will really happy and good luck!

Yours kindly

2019-03-15 11:57:42

Hi again. So I found a solution to this, If you agree with it oliver. If that's not a problem for you and if you own the api, You can send it to me in any way you like and I can set it up on our website which is not blocked by iran, nor it blocks iran. Please let me know if you agree with it so we can find a way to transfer the api. Thanks.

---
Co-founder of Sonorous Arts.
Check out Sonorous Arts on github: https://github.com/sonorous-arts/
my Discord: kianoosh.shakeri2#2988

2019-03-15 13:52:45

UltraLeetJ wrote:
oliver.edholm wrote:

Cool, that's awesome! Do you know remember name of that add-on?

I think it was called "describeIt"
there is info on that link. I think that one does OCR

Cool thanks for telling! I took a look, it required you to provide an API key to some cloud provider though, I guess that was too troublesome and because of that never took off hmm

2019-03-15 13:54:20

MasterOfDeath wrote:

Well, you could add that of course, and that would be really useful for many of us. OCR is sometimes not that accurate and if you could also implement that somehow, maybe it would improve, together with the image recognising feature. But as a starting, it's absolutely good, and if you keep developing it, we'll appreciate it. smile

Ok, OCR will definitely be added at some point. Right now I'm working on a CAPTCHA solving add-on which seems to be in a very dire need?

2019-03-15 13:55:25

baby wrote:

grate !. I can use in facebook!. especially when there are ! pictures  NVDA can tell what the picture is

Awesome! smile