Allows you to quickly read handwritten or printed paper Braille text using a smartphone or computer:

1. Take a photo, taking into account recommendations
2. Get the recognition results as:
• recognized text in braille characters,
• text in ordinary (flat-printed) letters,
• images with the signed meaning of each braille character.
3. Read, save or send the result to email.

By using the service, you agree to the User Agreement and Privacy Policy.

• teachers working with blind students,
• parents and relatives of blind people,
• any people who communicate with the blind and who need to read the text in Braille.

The service can be installed as an application on a smartphone, .

A pop-up "Install" button appears in the lower right corner. If you press it, the application will be installed on your smartphone.

An alternative way (for Chrome users) is to open the browser settings menu and select "Install Application".

Although you can use the service without registration, registration is required to fully work with the service. It's free. Why is it needed?

• First of all, you will be able to view previously uploaded texts. Since these are your personal texts, you have to protect them with a username and password;
• You can do default settings, so as not to choose the operating mode every time;
• You can revoke permission to publish previously uploaded materials after they have been uploaded and decrypted;
• Unregistered users cannot prohibit the public distribution of uploaded texts.

When registering, you can specify any username you like. However, we recommend that you specify a real E-mail address so that you can recover your password if you forget it.

In order for artificial intelligence-based systems to work, they need to be trained on a large number of examples. Initially, the neural network underlying Angelina Reader was trained on a set of Braille texts, published by Chinese researchers and on texts independently prepared by the developers of Angelina Reader. However, in order to make the quality of work better, additional training is needed on those images, which are uploaded by users of the service. In addition, since Angelina Reader owes its appearance to a data set, published in the public domain, we consider ourselves indebted to other researchers and would like to publish a publicly available data set.

If your text does not contain personal or confidential information, and you tick the "agree" box under the question of public access, You are contributing to the improvement of Braille recognition algorithms.

Of course, if you do not agree that the uploaded text should be seen by the developers, You can choose the "reject" option. In this case, no one will see the content of the text you uploaded.

To get a satisfactory result, you need to follow a few simple rules:

1. the light should fall from the (upper) side of the sheet farthest from you and give a contrasting image of points with light and shadow.
2. The sheet should be photographed from above.
3. The image scale is such that the A4 sheet occupies almost the entire frame.
4. Resolution of at least 1000 points vertically and horizontally (photos from most smartphones and cameras are suitable).

Right	Wrong (the light falls from above, from the side or from behind)	Wrong (it is necessary to photograph from above)

The system for recognition can be loaded with:

• when starting from a smartphone - a photo taken with the camera directly from the program;
• a photo taken in advance (a file with an image of the type *.jpg, etc.);
• file in Adobe Acrobat format (file type *.pdf, obtained, for example, from a scanner);
• archive in *.zip format containing images. To get an archive in Windows, select several files or a folder with files, right-click and select "send to - compressed zip folder";

Before uploading an image, you are prompted to set the text processing parameters:

• Consent to public access to the uploaded text or refusal from it.
• The language of the text. Determines the flat text in which the recognized Braille text will be translated.
• Auto-orientation. If this mode is enabled, the program itself determines how to rotate the sheet in the image correctly. The fact is that when you take a picture from top to bottom, the smartphone may choose the wrong orientation, the photo is turned upside down or turned on its side. The "auto-orientation" mode allows you to correctly recognize even inverted images. However, sometimes automatic orientation detection can fail. In this case, make sure that the photo is oriented correctly and disable the "auto-orientation" option. It is recommended to always select the "auto-orientation" option, except when using it gives incorrect results.
• Both sides. This is a mode that, based on the image of only one side of a sheet with double-sided printing, allows you to get text from both sides at once: front and back. However, the quality of recognition of the reverse side will be worse, therefore, it is recommended to enable this mode only when you need to quickly recognize the text for informational purposes. For a better result, we recommend taking a separate photo of each side and recognizing it separately, leaving the "both sides" mode off.

Registered users can set default values for all listed parameters in in the "settings" section.

Registered users can access the history of previously uploaded texts. If you need to go back to the previous page, you don't have to recognize it again. It is enough to find it in the list of previous results.

The last 5 documents are visible on the main page. The full list is available in the "history" section or via the "show all" link at the end of the short list.

Each entry in the history contains:

date and time when the image (picture) was uploaded name of the uploaded file the first 3 lines of the uploaded text icon that shows if public access is allowed or not .

Clicking on a line in the history opens recognition results in the same form as after loading the image.

Clicking on the lock icon allows you to change the decision about, is public access allowed to the image.

You can also navigate through the results from the history in the results form using the buttons and .

The recognition results are displayed on a separate page and include:

• image with letters of flat text signed under braille characters;
• recognized flat text;
• recognized braille text.

It an archive is loaded, all files in the archive will be displayed sequentially.

If the "both sides" option is enabled, two pages with results will be shown for each uploaded image: the front and back side.

You can use the buttons and to navigate between previously recognized images.

Using the "Send to E-mail" button, the result can be sent to e-mail.

To send to E-mail, click the "send to E-mail" button on the results page.

In the e-mail form you must enter the recipient's address.

You can specify the subject of the letter. If you leave this field empty, the subject of the email will be "Recognized braille name of the uploaded file"

You can also choose what to send: image, text or Braille text.

In addition, you can add a text comment that will be included in the letter.

If an error is detected, you can send an email or a copy of it to the developers by checking the appropriate box. In this case, you will need to fill in the "comment" field. There you need to describe the problem you want to report.

The system allows you to configure one or more recipient addresses for quick selection using the button to the right of the destination address field. The list of addresses to select is filled in in the "settings" section.

Also in the "settings" section, you can configure the default values for the mail sending mode.

The "Settings" section allows you to change

• password
• user's display name
• default settings for recognition modes and mail sending modes
• list of the most frequently used recipient addresses