The call succeeded for the first 5 times but returned garbled ocr text. It took about 30 seconds per page. I am using the 2.2 version. I am also not sure if I need to set the english language and/or set the resources property (which no longer seems to exist in version 2.2)
The call succeeded for the first 5 times but returned garbled ocr text. It took about 30 seconds per page. I am using the 2.2 version. I am also not sure if I need to set the english language and/or set the resources property (which no longer seems to exist in version 2.2)
Since the release of Aspose.OCR for .NET 2.2.0, you do not need to set the language as well as the resource property. The aforesaid release has embedded the resources to perform the OCR operation for English language (default language) into the Aspose.OCR assembly.
Your code is correct as we have evaluated the case with a few sample images of our own. Unfortunately, we are unable to observe the said problem. We suspect that the problem could be more related to the sample itself, therefore we would request you to please share the sample image for further investigation.
Thank you for providing the sample image. By checking the provided screenshots, we have noticed that you are setting the OcrConfig.Whitelist property that is causing the IndexOutOfRange error for some unknown reason. We have logged this problem in our bug tracking system under the ticket OCR-33918 for further investigation and correction purposes. Please spare us little time to properly analyze the scenario to isolate the problem cause. In the meanwhile, we will keep you posted with updates in this regard.
Please note, if you do not set the aforesaid property, you can avoid the exception and the process will not hang as well, so you do not need to downgrade the API version for this reason.
I removed the Whitelist setting and it did not crash. Unfortunately, it took 20 seconds to process a 300 dpi image that has only 5 upper case test words on it. It got a poor text result as well. See the attachment. I am surely doing something wrong. Any suggestions would be greatly appreciated.
We have logged the problem of poor recognition results in our bug tracking system under the ticket OCR-33921 for further investigation. Please allow us some time to properly analyze the scenario to isolate the problem cause, and to provide the fix at earliest possible. In the meanwhile, we will keep you posted with updates in this regard.