I would like to know how reliable TESSERACT is compared to other OCRs
I'm doing a configuration of an instance, and with the tesseract I had one of the best PDF readings, with and without images. While AZURE was not as accurate with the other documents.
Can anyone share their experience on this.
Hi @Gabriel O ,
The most reliable OCR is Tesseract, when you're using the Tesseract you can see the more accurate group creations in the training period so that will helps to increase the STP in production.
I have tried that, 100 documents were used to create the instance by using Tesseract and Azure OCRs
Tesseract OCR instance created the 80 groups and Azure created the 65 groups. the best accuracy for classification for Tesseract OCR.
Note: You have to think about document quality as well. For good-quality documents, you can use Tesseract. For scanned documents please use the Azure.