Identifying check box selection using OCR

Aspose.OMR Product Family
1

How to identify check box selection on image using OCR

2

@pkumar1826

Is it possible that you can please share a sample image? We need to investigate it and will log an investigation ticket in our issue management system.

3

Thanks for the quick reply

Here is the image with check box selected.
CheckBoxImage.jpg (3.2 KB)

Actually this section is part of by bigger document which has few tables with data.

4

@asad.ali Thank you for looking into this issue. Can you please provide a timeframe to which we’ll get an update on this, as we are currently doing PI planning and need to understand if this feature is supported.

5

@pkumar1826

We have opened the following new ticket(s) in our internal issue tracking system and will deliver their fixes according to the terms mentioned in Free Support Policies.

Issue ID(s): OCRNET-872

Can you please provide a timeframe to which we’ll get an update on this, as we are currently doing PI planning and need to understand if this feature is supported.

@lholmes9608

We have just logged a ticket and will continue investigating the feasibility. As soon as it is investigated, we will be able to share some timeframe for its resolution with you. Please be patient and spare us some time.

1 Like
6

Hello, @pkumar1826, @lholmes9608

For checkboxes and bubbles, we have a separate solution - Aspose.OMR (Optical-Mark-Recognition). Example

Recognition stages are different:

  1. Aspose OMR generates a template with a special reference points (black squares around corners) and an .omr file with embedded recognition pattern, unique to this template.
  2. Template copies are printed, filled and scanned.
  3. Both a scan of the completed template and the .omr file with the unique pattern are used for recognition.

If this process is acceptable, - we have an online tool to create .omr patterns from existing templates.sample.png (254.6 KB)
And a free plugin for handwriting text recognition.

7

Thanks for the reply @nikita.korobeynikov

But in our case, the check box is part of a form which has other fields that needs to be captured as well.
Can you provide any example for this case

8

@pkumar1826
We have an array of different input elements, which can be combined into more complex features.

Do you require OCR and OMR features?
Can you share a sample of whole form?

9

@nikita.korobeynikov

Here is the document we are trying to read.

Also please mention which DetectAreasMode is suitable for this kind of form

ocrDoc.jpg (179.0 KB)

Thanks

10

@pkumar1826

We are checking it and will get back to you shortly.

11

Hello, @pkumar1826

If I am understanding correctly - hard copies already filled and scanned, meaning it is not possible to update document and process again (filling or scanning). In this case, - there are no features for this variant yet.

At the current moment, Aspose.OMR works only with template generated by itself (hence a big element generation API) and require template to include additional reference points (see black squares in corners of the sample. This is done to compensate any changes in the scanned images.

Aspose.OCR can work with scans as is, but does not have an API to process marks.

It is an interesting and important case, and we will include features for it in the development queue and, if you’re interested, notify you in this thread upon completion.
As free support task it will be in a free queue and can be moved by a Paid consulting tasks.