I have a UWP app and I need to OCR a document.
I use the API of Windows.Media.Ocr (OcrEngine.RecognizeAsync(SoftwareBitmap))
In some cases, the recognized text is not perfect, especially for underlined text, even for "normal" printed plain English text, coming from a scanned book.
Is there any way to configure the API or to run some operations (e.g. preprocessing of images) to get a better recognition quality?
Thanks!
Is there any way to configure the API or to run some operations (e.g. preprocessing of images) to get a better recognition quality?
I'm afraid currently there is no such property for OcrEngine, if you do want this feature, please feel free post your requirement with windows feed backhub. There is no limit for the image type, but higher resolution image could be better recognize. For image type we suggest use png or bmp.