I had been testing and evaluating few Aspose dlls- PDF, Words, Email and OCR. Most of them worked well for our needs, though the OCR was very problematic. It returned me strange text results even when trying to extract text from real simple and clear text images. I am wondering what could be the issue and if it could be something related to configurations. I used the Aspose.OCR 2.2.0.0 dll for .Net. I am working with a .net 4.5 platform.
Hi Fadi,
Could you please provide us a sample image for testing & investigation?
Hi Babar,
Here are two images for testing:
http://booksprung.com/wp-content/uploads/2009/11/ocr-test-source-material.jpg
https://jurnsearch.files.wordpress.com/2009/07/ocr-test.jpg
Thanks
Fadi
Hi Fadi,
Thank you for providing the samples for our testing.
We have evaluated your presented scenario while using the latest version of Aspose.OCR for .NET 2.3.0, and we are able to observe incorrect recognition results for both samples. The most probable reason for such a low recognition accuracy is the quality of the images (72 & 96 DPI). Please note, Aspose.OCR APIs works well with high resolution images with at least 300 DPI.