When using Nuance to do full page OCR in Ephesoft for Linux, incorrect coordinates are coming for some of the words in HOCR.xml. (Zero value is getting populated for the coordinates.)
Duplicate entries are present for some of the words in HOCR.xml consecutively.
Download the file at the following link: NuanceOCR.zip
STEPS TO APPLY THE SOLUTION:
- Extract the NuanceOCR.zip to a temporary folder.
- Stop the Ephesoft server.
- Take back-up of NuanceOCR executable file located at /opt/Ephesoft/Application/native/Nuance/*.
- Copy the extracted NuanceOCR executable file to /opt/Ephesoft/Application/native/Nuance/*.
- Change permissions of the executable file using the following command:
chmod 777 /opt/Ephesoft/Application/native/Nuance *
- Restart the server.