Guidance Required for Capturing OMR Sheets via Webcam in C# ASP.NET MVC

I am currently developing a solution in C# ASP.NET MVC to capture and process OMR sheets using a webcam. While I have successfully uploaded and evaluated images, I need assistance with real-time webcam-based OMR sheet detection.

Could you provide guidance on the following:

  1. Best practices for capturing OMR sheets via a webcam in an ASP.NET MVC application.
  2. Recommended resolution, lighting, and angle for optimal OMR detection.
  3. Any SDKs, APIs, or third-party libraries you suggest for real-time image processing in C#.
  4. Sample code or implementation references for webcam-based OMR processing.

I would appreciate any documentation, examples, or recommendations that can help implement this feature efficiently.

Looking forward to your response.

Best regards,
Veera

@StanlyIGX

Can you please specify which specific issues you are facing with real-time webcam-based OMR sheet detection in your C# ASP.NET MVC application?

I am developing a C# ASP.NET MVC application for capturing and processing OMR sheets using a webcam. While uploading and evaluating images works well, I am facing issues with real-time webcam-based detection.

Issues Faced:

  1. Sometimes, the webcam captures an image that includes areas outside the OMR sheet, which causes the evaluation to fail.
  2. When an invalid or incorrectly framed image is captured, the system throws a 500 error instead of handling it gracefully.
  3. Need suggestions on how to pre-process the captured image (e.g., cropping, edge detection) before evaluation to avoid incorrect submissions.

Request for Assistance:

  • Recommended approach to ensure only the OMR sheet is captured using a webcam.( Very important)
  • Best practices to handle incorrect or incomplete OMR captures and prevent errors.
  • Any built-in validation techniques or SDK recommendations for improving detection accuracy in real-time.
  • Sample code or documentation for optimizing real-time OMR sheet detection.

Your insights would be highly valuable in resolving these issues. Looking forward to your guidance.

Best regards,
Veera

Hello, Veera.

I have registered a support ticket for this thread. Number is - OMRNET-1107.

  1. There is margin of space that reserved for this occurrence. Can you share an image that were failing during template recognition?

  2. Based on thread tags, I am assuming that .NET library is used for template generation. “500 error” - does it mean that an OMR Cloud solution were used and are throwing a 500 http status error?
    If it is - samples of incorrectly framed images, requests and responses would help to resolve this issue.

  3. Technical details for next questions would help to provide a more precise suggestion:

  • Is this webcam a standard model and have predefined characteristics that we can investigate, or it is unknown and will be only be available at runtime?

  • On previous topics, there is a mention of Freehand text recognition. Is this a required feature? There is an array of techniques that produce a more precise result in identifying person perfoming a test.

  • On what environment would code run and what computing capabilities would be available?