Android Question Capture image with camera and recognize image points

[rscheel]

I need to capture an image of an answer sheet and then process it and recognize the good and bad is it possible to do that?

 

[DonManfred]

I don´t think there is a library for this.

You need to analyze the picture by yourself.

I think it will be an complicated task.

 

[rscheel]

I don´t think there is a library for this.

You need to analyze the picture by yourself.

I think it will be an complicated task.

There is an application on iOS that does that, I have to make an app on Android and iOS.

 

[DonManfred]

There is an application on iOS that does that

It is NOT relevant!

We are talking about ANDROID here!

Even if there are HUNDRETS of such apps... How will it help YOU building yours? Right; it is not a help....

 

[JordiCP]

OpenCV can help you a bit.

It has a lot of classes and methods to extract features from an image that can save you a lot of work. BUT anyway you'll have to develop your own 'higher level' algorithm (using OpenCV or directly analyzing the image) to get what you want.

I think it will be an complicated task.

Agree. It's not straightforward.

 

[rscheel]

I figured, when I saw how data capture should work, it was not easy at all.

 

[Cableguy]

From the videos shown, I see "pattern" comparison, specially on the second example where we can clearly see "alignment" squares to help the digitising.
So I guess that, complicated, yes, specially finding the "test" boundaries, then it's a matter of creating a grid, and compare it to a master pattern.
Of course, this is my theory...

 

[rscheel]

The same application is developed for android.

https://play.google.com/store/apps/details?id=com.zipgradellc.android.zipgrade

This interesting topic, but the bad thing is that the client wants to develop similar App in a period of 2 months, for Android, iOS and also develop the web system to manage the data, it is a lot of work for me, and only the first stage of the project.

 

[NJDude]

Here is a clue, to do this you need a "Optical Mark Recognition" (OMR) library for B4A.

 

[rscheel]

In case someone is encouraged to give you a look, I found this information, maybe with google vision you can achieve this.

https://github.com/rbaron/omr

 

[JordiCP]

The python script uses OpenCV (all the cv2.xxxx). With a similar syntax, all those methods are available in OpenCvforB4A.
I'm not sure if that script is tuned for a specific sheet, but it seems a good starting point.

PM me if you want the lib (it's free )

 

[rscheel]

The python script uses OpenCV (all the cv2.xxxx). With a similar syntax, all those methods are available in OpenCvforB4A.
I'm not sure if that script is tuned for a specific sheet, but it seems a good starting point.

PM me if you want the lib (it's free )

Tu trabajo con esa librería es grandioso, muy bueno, voy a probarla pero igual voy a necesitar mucho tiempo ya que estoy haciendo varios proyectos y todos son para ayer, y me salio este proyecto nuevo el cual es muy ambicioso para el corto plazo de desarrollo, lamentablemente no me da el tiempo, pero siempre es bueno indagar en estos temas para el futuro.

Your work with this bookstore is great, very good, I'm going to try it but I'm going to need a lot of time since I'm doing several projects and all are for yesterday, and I left this new project which is very ambitious for the short term development , unfortunately I do not give the time, but it is always good to inquire into these issues for the future..