Evaluation of Template Editor and Aspose.OCR

We are continuing our review of the Aspose.OCR SDK for OMR sheet processing but have run into a few more problems –



• When a field has multi-character values (such as “2015”) and it is not marked, only a single blank is output instead of 4 blanks, requiring us to put another layer of code on the results from the Template.



• We occasionally have to use “extended grids” when one grid ends and additional values appear in a related grid. For example, suppose we have 3 equally spaced bubbles and then 2 more bubbles with different spacing. If that were treated as 2 grids, there would be 2 output responses, but only one is allowed logically. Again, another layer of code is necessary to look at related grids and use the single non-blank value for output.



• Apparently, it is not possible to define the direction of recognition for grids and we have found it sometimes necessary to reverse the output order of information in a field.



• We have been unable to define fields with more than 10 values. Thus, a field with 12 values like months of the year is not possible.



The upshot of this is that we cannot always use the Template Editor to define the output for a recognition batch. We have to provide additional coding to manipulate the raw output. Are there any plans to make the Template Editor more like OMR editors with commonly used grid definitions?

Hi Dan,

Thank you for your inquiry.

We are looking into your requirements. We will update you soon.
Hi Dan,

Following are the answers to your questions:

Point #1:
If “2015” is one element then you have to use four elements for “2015”. In this way you will get all values separately in return.

Point #2:
In this case we need the sample OMR template and the scanned OMR sheet image for better understanding.

Point #3:
This requires more information. Please clear that if a Grid element has values from 1 to 10, do you want to read value 10 first and so on up till value 1?

Point #4:
If we use Numeric values then YES, 0 to 9 makes 10 values and we cannot define more than 10. On the other hand if we choose Alphabetic then we can go from A to Z which makes it more than 10 values. However, we have logged this requirement into you system with ID OCR-34236. Our product team will look into it. We will update you accordingly.

Hope the above information helps. Feel free to contact us in case you have further comments or queries.

Ikram,



These all essentially have to do with how much the Template Editor can do and how much translation is done before the data are output.



For example, if we have a bubble for “2014”, one for “2015”, and one for “2016”, do we output the 4 associated characters or simply a “1”, “2”, or “3” and later interpret the meaning. Another example is selection of one of 50 US states which represents several bubble definition problems.



We realize that this is an SDK that can’t be expected to do everything and that additional code will be necessary. Our decision is whether we want to write an enhanced Template Editor or purchase a product with a User Interface more like the standard OMR template editor.



That is our decision! Thanks!

Hi Dan,

Note that OMR Template Editor is to facilitate the OMR process. If you want to recognize “2014” or “2015” then you have to use four elements for “2015”. In this way you will get all 4 values separately in return. You will not get 4 associated characters rather you will get “0”, “1”, “2” or “3”.

Further to update you that our product team is looking into the ticket OCR-34236. We will update you accordingly via this thread.

Hi Dan,

Thank you for your patience.

Please note that in order to work on issue “defining fields with more than 10 values” we need sample templates that you are using. Sharing such templates will help us in batter understanding the issue you are facing.

Site wouldn’t allow attached file.

An example would be a single column to accept the numbers or characters “01” to “12”. If not marked, two spaces should be in the output, but current version only inserts a single space, disturbing the structure of the output record.

Another example would be 5 alpha columns of 10 each for the 2-character codes of the 50 US states (AL, AK, …, VA, WY). The columns need to be dependent so that when an item in one of t he 5 columns is marked, only it will be output, not a character for each of the 5 columns.

Thanks!

Dan Eckard

Hi Dan,

Please login to your Aspose account, ZIP the sample files and then try to attach/upload with the reply post. In this way you will be able to send us the sample files.

Sorry, the zipped .amr file was 10MB and took a while to upload. Here it is. See the COURSE CODE / Course columns with the months, Jan - Dec.



Dan

Hi Dan,

Thank you for the sample template.

Please note that we will now proceed with the investigation and will update you accordingly.

Hi Dan,

Thank you for your patience.

Please find the updated OMR template (ExamUpdate.amr) attached with the reply. We have made some corrections in the template mainly replacing Grid elements with ChoiceBox elements to correctly recognize marked answers.

Furthermore you may use Grid element to recognize ID fields, Name fields and similar data.

Can I get my trial reactivated so I can test it?

Hi Dan,

Thank you for your inquiry.

This is to update you that one time reactivation of 30 day trial is possible. Please post your inquiry regarding reactivation on the Aspose.Purchase support forum. The Aspose.Purchase support team will update you accordingly.

The issues you have found earlier (filed as ) have been fixed in this Aspose.Words for JasperReports 18.3 update.