Topics

Automatic alt text #addonrelease

Oliver Edholm
 
Edited

Hi everyone!
 
This is an add on 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.
 
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback! 
 
The shortcut to interpret an image is control+NVDA+A.
 
The add on can be downloaded at http://www.screenreader.ai/
 
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.

Rui Fontes
 

Attention!
NVDA+Control+A is a NVDA keystroke to select a Braille display...

Rui Fontes


Às 15:38 de 07/03/2019, Oliver Edholm escreveu:

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.

 

Hi Oliver,

Thanks for this add-on.

 

After reviewing this add-on, it appears there are some serious items you may wish to take a look at, including specifying copyright, correct manifest strings, and user experience issues. See below for comments.

 

For resident reviewers, basic review results are as follows:

 

  • License and copyright: fail
  • Documentation: fail
  • User experience: fail
  • Security: no comment

 

Comments:

 

  1. Can you let us know under what license this add-on is copyrighted? We are being careful about this. Perhaps you can state this at the top of the source code file (in comments).
  2. About readme: I see that you have built the add-on directly from the add-on template, as evidenced by the readme content. I advise not doing this in the future – the template should be used as a guide, not a tool for building the add-on.
  3. User experience isn’t working – perhaps it’s my computer. When I try to analyze a test image, it comes back and describes what that image is instead of giving me a text that’s inside the image in question.
  4. The command Control+NVDA+A conflicts with a command to choose braille displays.

 

An important note about add-on name as shown in the manifest: add-on name is the identifier, and the description is what users will see in Add-ons Manager. Perhaps try using something like:

 

name = autoAltText

 

The review policy states that we cannot proceed with other reviews until license and copyright problem is resolved. For this reason, I would like to suggest resolving this as soon as possible. Thanks for understanding.

 

And with that, I’m out of office.

Cheers,

Joseph

 

 

From: nvda-addons@nvda-addons.groups.io <nvda-addons@nvda-addons.groups.io> On Behalf Of Oliver Edholm
Sent: Thursday, March 7, 2019 7:39 AM
To: nvda-addons@nvda-addons.groups.io
Subject: [nvda-addons] Automatic alt text #addonrelease

 

[Edited Message Follows]

Hi everyone!

 

This is an add on 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.

 

I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback! 

 

The shortcut to interpret an image is control+NVDA+A.

 

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

 

Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.

Rui Fontes
 

Hello Joseph!

You wrote:
3. User experience isn’t working – perhaps it’s my computer. When I try
to analyze a test image, it comes back and describes what that image
is instead of giving me a text that’s inside the image in question.
Is not that the purpose of this addon?

Rui Fontes

 

Hi,
At least description is part of it, but what I was looking for is a common case people would use this add-on for: text recognition.
Cheers,
Joseph

-----Original Message-----
From: nvda-addons@nvda-addons.groups.io <nvda-addons@nvda-addons.groups.io> On Behalf Of Rui Fontes
Sent: Thursday, March 7, 2019 8:10 AM
To: nvda-addons@nvda-addons.groups.io
Subject: Re: [nvda-addons] Automatic alt text #addonrelease

Hello Joseph!

You wrote:
3. User experience isn’t working – perhaps it’s my computer. When I try
to analyze a test image, it comes back and describes what that image
is instead of giving me a text that’s inside the image in question.
Is not that the purpose of this addon?

Rui Fontes

James Scholes
 

Thanks for your work on this, I can imagine this being useful for a lot of people. Some initial thoughts to add to what Joseph has already said:

1. Is the current add-on name appropriate? The add-on is not automatic, users need to activate it manually. Plus, many users don't know what alt text is. Perhaps "Image Descriptions", "Image Describer" or something of that nature would be better.
2. You don't need to define an __init__ method inside your GlobalPlugin if all it does is call the superclass's implementation. If you plan to extend your custom __init__ method in the future to add some functionality, please update the signature as follows:

def __init__(self, *args, **kwargs):
super(globalPluginHandler.GlobalPlugin, self).__init__(*args, **kwargs)

Regards,

James Scholes
https://twitter.com/JamesScholes

Timothy
 



Isn't this a duplicate of DescribeIt, a project of the Indian Institute of Technology?

Timothy

----- Original Message -----
Sent: Thursday, March 7, 2019 10:38 AM
Subject: [nvda-addons] Automatic alt text #addonrelease

[Edited Message Follows]

Hi everyone!
 
This is an add on 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.
 
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback! 
 
The shortcut to interpret an image is control+NVDA+A.
 
The add on can be downloaded at http://www.screenreader.ai/
 
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.

Nikola Jovic
 

Make sure that you are not using screen curtain, as of course the
addon wont be able to see anything.

On 3/7/19, Michal Suchý <suchy.michal@...> wrote:
Hello. I tryed this addon and everywhere Nvda says black, darkness, white,
sky.



 

Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up the good work!
I'd like so much to get into a project to translate this one, but probably it won't be possible since I don't know well how to use GitHub.
Anyway, I hope to hear more of it!


Cheers,
Marcio AKA Starboy
Follow or add me on Facebook

Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
 
This is an add on 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.
 
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback! 
 
The shortcut to interpret an image is control+NVDA+A.
 
The add on can be downloaded at http://www.screenreader.ai/
 
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.

Rui Fontes
 

Do you mean the addon messages or the description itself?

Rui Fontes


Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:

Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up the good work!
I'd like so much to get into a project to translate this one, but probably it won't be possible since I don't know well how to use GitHub.
Anyway, I hope to hear more of it!
Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>
Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.

 

Actually, when I talk about translating, I do it taking all aspects into account (i.e. description, messages and maybe even the add-on's name).


Cheers,
Marcio AKA Starboy
Follow or add me on Facebook

Em 07/03/2019 20:53, Rui Fontes escreveu:

Do you mean the addon messages or the description itself?

Rui Fontes


Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:
Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up the good work!
I'd like so much to get into a project to translate this one, but probably it won't be possible since I don't know well how to use GitHub.
Anyway, I hope to hear more of it!


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.







Timothy
 



The image description comes from an online service, so it would be up to that service to provide description in other languages.

Timothy

----- Original Message -----
Sent: Thursday, March 7, 2019 7:02 PM
Subject: Re: [nvda-addons] Automatic alt text #addonrelease

Actually, when I talk about translating, I do it taking all aspects into account (i.e. description, messages and maybe even the add-on's name).


Cheers,
Marcio AKA Starboy
Follow or add me on Facebook

Em 07/03/2019 20:53, Rui Fontes escreveu:
Do you mean the addon messages or the description itself?

Rui Fontes


Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:
Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up the good work!
I'd like so much to get into a project to translate this one, but probably it won't be possible since I don't know well how to use GitHub.
Anyway, I hope to hear more of it!


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.







 

So isn't there a way to translate these messages?
I mean, if we could intercept the messages then put them into a kind of variable, it'd  easy to just translate the result from that variable and replace with the new one.

Cheers,
Marcio AKA Starboy
Follow or add me on Facebook

Em 07/03/2019 21:24, Timothy via Groups.Io escreveu:



The image description comes from an online service, so it would be up to that service to provide description in other languages.

Timothy

----- Original Message -----
Sent: Thursday, March 7, 2019 7:02 PM
Subject: Re: [nvda-addons] Automatic alt text #addonrelease

Actually, when I talk about translating, I do it taking all aspects into account (i.e. description, messages and maybe even the add-on's name).


Cheers,
Marcio AKA Starboy
Follow or add me on Facebook

Em 07/03/2019 20:53, Rui Fontes escreveu:
Do you mean the addon messages or the description itself?

Rui Fontes


Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:
Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up the good work!
I'd like so much to get into a project to translate this one, but probably it won't be possible since I don't know well how to use GitHub.
Anyway, I hope to hear more of it!


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.








Locutor Antonio Cezar
 

Or use an add-on that translates everything spoken by NVDA, such as Translate, for example ...



Please, sign our petition in the link below, share with your friends, and collaborate with us! In favor of Functional Braille on Android!

https://secure.avaaz.org/po/petition/Google_Inc_Talkback_Development_Team_Google_Disability_Support_Team_PRECISAMOS_DO_BRAILLE_NO_ANDROID_TAMBEM/

Timothy via Groups.Io escreveu:



The image description comes from an online service, so it would be up to that service to provide description in other languages.

Timothy

----- Original Message -----
Sent: Thursday, March 7, 2019 7:02 PM
Subject: Re: [nvda-addons] Automatic alt text #addonrelease

Actually, when I talk about translating, I do it taking all aspects into account (i.e. description, messages and maybe even the add-on's name).


Cheers,
Marcio AKA Starboy
Follow or add me on Facebook

Em 07/03/2019 20:53, Rui Fontes escreveu:
Do you mean the addon messages or the description itself?

Rui Fontes


Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:
Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up the good work!
I'd like so much to get into a project to translate this one, but probably it won't be possible since I don't know well how to use GitHub.
Anyway, I hope to hear more of it!


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.







M.R.
 

Could you place the description in the clipboard?
It would be very convenient.
And i need a possibility to change a hot key.
Thanks!!!


08.03.2019 5:53, Locutor Antonio Cezar пишет:

Or use an add-on that translates everything spoken by NVDA, such as
Translate, for example ...


------------------------------------------------------------------------

Please, sign our petition in the link below, share with your friends,
and collaborate with us! In favor of Functional Braille on Android!

https://secure.avaaz.org/po/petition/Google_Inc_Talkback_Development_Team_Google_Disability_Support_Team_PRECISAMOS_DO_BRAILLE_NO_ANDROID_TAMBEM/

Timothy via Groups.Io escreveu:


The image description comes from an online service, so it would be up
to that service to provide description in other languages.

Timothy

----- Original Message -----
*From:* marcio via Groups.Io
<mailto:marcinhorj21=yahoo.com.br@groups.io>
*To:* nvda-addons@nvda-addons.groups.io
<mailto:nvda-addons@nvda-addons.groups.io>
*Sent:* Thursday, March 7, 2019 7:02 PM
*Subject:* Re: [nvda-addons] Automatic alt text #addonrelease

Actually, when I talk about translating, I do it taking all
aspects into account (i.e. description, messages and maybe even
the add-on's name).


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 20:53, Rui Fontes escreveu:
Do you mean the addon messages or the description itself?

Rui Fontes


Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:
Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up
the good work!
I'd like so much to get into a project to translate this one,
but probably it won't be possible since I don't know well how to
use GitHub.
Anyway, I hope to hear more of it!


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with
NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every
time you analyze an image, this is because downloading all the
computer vision libraries required locally would take way too
much storage.


Abdel
 

Hi,

Thank you for this very useful add-on.

It would be interesting if such an add-on is present on our community website after following these instructions.

  • Follow the recommendations described by Joseph in his review;
  • Delete the __init__ method since it doesn't add anything as suggested by James;
  • Define a category so that the gesture for the script can be editable.


Otherwise, in terms of compatibility with Python 3, as well as the ability to copy the description, this will be relatively simple to add, after following the above advice.

Thanks again for this great work.

Kind regards,
Abdel.


Le 08/03/2019 à 04:42, M.R. a écrit :
Could you place the description in the clipboard?
It would be very convenient.
And i need a possibility to change a hot key.
Thanks!!!


08.03.2019 5:53, Locutor Antonio Cezar пишет:
Or use an add-on that translates everything spoken by NVDA, such as
Translate, for example ...


------------------------------------------------------------------------

Please, sign our petition in the link below, share with your friends,
and collaborate with us! In favor of Functional Braille on Android!

https://secure.avaaz.org/po/petition/Google_Inc_Talkback_Development_Team_Google_Disability_Support_Team_PRECISAMOS_DO_BRAILLE_NO_ANDROID_TAMBEM/

Timothy via Groups.Io escreveu:


The image description comes from an online service, so it would be up
to that service to provide description in other languages.

Timothy

    ----- Original Message -----
    *From:* marcio via Groups.Io
    <mailto:marcinhorj21@...>
    *To:* nvda-addons@nvda-addons.groups.io
    <mailto:nvda-addons@nvda-addons.groups.io>
    *Sent:* Thursday, March 7, 2019 7:02 PM
    *Subject:* Re: [nvda-addons] Automatic alt text #addonrelease

    Actually, when I talk about translating, I do it taking all
    aspects into account (i.e. description, messages and maybe even
    the add-on's name).


    Cheers,
    Marcio AKA /Starboy/
    Follow or add me on Facebook <https://facebook.com/firirinfonfon>

    Em 07/03/2019 20:53, Rui Fontes escreveu:
    Do you mean the addon messages or the description itself?

    Rui Fontes


    Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:
    Hi,
    I just tested this add-on.
    First, congratulations for the job and thanks for the initiative!
    For an 1.0 version it's actually running quite well in my opinion.
    I can see just a brilliant future for this add-on, so keep up
    the good work!
    I'd like so much to get into a project to translate this one,
    but probably it won't be possible since I don't know well how to
    use GitHub.
    Anyway, I hope to hear more of it!


    Cheers,
    Marcio AKA /Starboy/
    Follow or add me on Facebook <https://facebook.com/firirinfonfon>

    Em 07/03/2019 12:38, Oliver Edholm escreveu:

    [Edited Message Follows]

    Hi everyone!
    This is an add on 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.
    I'm not visually impaired and have limited experience with
    NVDA, would love all kinds of feedback!
    The shortcut to interpret an image is control+NVDA+A.
    The add on can be downloaded at http://www.screenreader.ai/
    Note that the add on sends a request to a REST endpoint every
    time you analyze an image, this is because downloading all the
    computer vision libraries required locally would take way too
    much storage.













Brian's Mail list account
 

I wonder how it copes with the various Cap Chars around?
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Michal Suchý" <suchy.michal@...>
To: <nvda-addons@nvda-addons.groups.io>
Sent: Thursday, March 07, 2019 10:06 PM
Subject: Re: [nvda-addons] Automatic alt text #addonrelease


Hello. I tryed this addon and everywhere Nvda says black, darkness, white, sky.

Brian's Mail list account
 

I'd imagine that would take far too long as you may need to send it to a translate service when if you knew you were running in a certain language the service itself would or should realise it and send the text in that language.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "marcio via Groups.Io" <marcinhorj21=yahoo.com.br@groups.io>
To: <nvda-addons@nvda-addons.groups.io>
Sent: Friday, March 08, 2019 12:29 AM
Subject: Re: [nvda-addons] Automatic alt text #addonrelease


So isn't there a way to translate these messages?
I mean, if we could intercept the messages then put them into a kind of
variable, it'd easy to just translate the result from that variable and
replace with the new one.

Cheers,
Marcio
AKA /Starboy/


Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 21:24, Timothy via Groups.Io escreveu:


The image description comes from an online service, so it would be up
to that service to provide description in other languages.

Timothy

----- Original Message -----
*From:* marcio via Groups.Io
<mailto:marcinhorj21=yahoo.com.br@groups.io>
*To:* nvda-addons@nvda-addons.groups.io
<mailto:nvda-addons@nvda-addons.groups.io>
*Sent:* Thursday, March 7, 2019 7:02 PM
*Subject:* Re: [nvda-addons] Automatic alt text #addonrelease

Actually, when I talk about translating, I do it taking all
aspects into account (i.e. description, messages and maybe even
the add-on's name).


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 20:53, Rui Fontes escreveu:
Do you mean the addon messages or the description itself?

Rui Fontes


Às 22:48 de 07/03/2019, marcio via Groups.Io escreveu:
Hi,
I just tested this add-on.
First, congratulations for the job and thanks for the initiative!
For an 1.0 version it's actually running quite well in my opinion.
I can see just a brilliant future for this add-on, so keep up
the good work!
I'd like so much to get into a project to translate this one,
but probably it won't be possible since I don't know well how to
use GitHub.
Anyway, I hope to hear more of it!


Cheers,
Marcio AKA /Starboy/
Follow or add me on Facebook <https://facebook.com/firirinfonfon>

Em 07/03/2019 12:38, Oliver Edholm escreveu:

[Edited Message Follows]

Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with
NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every
time you analyze an image, this is because downloading all the
computer vision libraries required locally would take way too
much storage.




Brian's Mail list account
 

I think that was already mentioned. but I was just doing my brain in trying to think of a key combination that did not clash, and have thus far failed, which is a little annoying.
Brian

bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.

----- Original Message -----
From: "Rui Fontes" <rui.fontes@...>
To: <nvda-addons@nvda-addons.groups.io>
Sent: Thursday, March 07, 2019 3:48 PM
Subject: Re: [nvda-addons] Automatic alt text #addonrelease


Attention!
NVDA+Control+A is a NVDA keystroke to select a Braille display...

Rui Fontes


Às 15:38 de 07/03/2019, Oliver Edholm escreveu:
Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.

Martina Letochová
 

What about NVDA+Control+I? I as for image.
Martina

On 8/3/2019 11:10, Brian's Mail list account via Groups.Io wrote:
I think that was already mentioned. but I was just doing my brain in trying to think of a key combination that did not clash, and have thus far failed, which is a  little annoying.
Brian
bglists@...
Sent via blueyonder.
Please address personal E-mail to:-
briang1@..., putting 'Brian Gaff'
in the display name field.
----- Original Message ----- From: "Rui Fontes" <rui.fontes@...>
To: <nvda-addons@nvda-addons.groups.io>
Sent: Thursday, March 07, 2019 3:48 PM
Subject: Re: [nvda-addons] Automatic alt text #addonrelease

Attention!
NVDA+Control+A is a NVDA keystroke to select a Braille display...

Rui Fontes


Às 15:38 de 07/03/2019, Oliver Edholm escreveu:
Hi everyone!
This is an add on 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.
I'm not visually impaired and have limited experience with NVDA, would love all kinds of feedback!
The shortcut to interpret an image is control+NVDA+A.
The add on can be downloaded at http://www.screenreader.ai/
Note that the add on sends a request to a REST endpoint every time you analyze an image, this is because downloading all the computer vision libraries required locally would take way too much storage.