FineReader Engine recognition process taking too much time

Symptoms

Processing and Export steps for documents with a high page count are taking more time than expected

 

Resolution

Some settings applied may artificially reduce the processing capacities of the working machine to the minimum amount of CPU cores used for processing.

Using IMultiProcessingParams::put_MultiProcessingMode( MPM_Sequential ) will force the Engine to process only one page at a time resulting in extended Processing and Export timers, this can be diagnosed by reviewing the logs. 

As can be seen in the sample:

50800 :16:16:05.025, 166 ms, IFRPage::DetectResolution( (IObjectsExtractionParams*)0x612a6e4, 50, 600, (int*)0xaff628 )
50800 :16:16:05.192, 434531 ms, IFRDocument::Process( (IDocumentProcessingParams*)0x4b87224 )
50800 :16:23:19.725, 0 ms, IPrepareImageMode::get_Rotation( (RotationTypeEnum*)0xaff5d0 )
50800 :16:23:19.725, 0 ms, IDocumentProcessingParams::get_SynthesisParamsForDocument( (ISynthesisParamsForDocument*__ptr64*)0xaff660 )
50800 :16:23:19.726, 3616 ms, IFRDocument::Synthesize( (ISynthesisParamsForDocument*)0x3af39394 )
50800 :16:23:23.343, 5 ms, IEngine::CreatePDFExportParams( (IPDFExportParams*__ptr64*)0xaff628 )
50800 :16:23:23.348, 0 ms, IPDFExportParams::put_MRCMode( MRC_Disable )
50800 :16:23:23.348, 0 ms, IPDFExportParams::put_TextExportMode( PEM_ImageOnText )
50800 :16:23:23.349, 0 ms, IPDFExportParams::get_PDFFeatures( (IPDFExportFeatures*__ptr64*)0xaff620 )
50800 :16:23:23.349, 0 ms, IPDFExportFeatures::put_WriteTaggedPDF( TSPV_Yes )
50800 :16:23:23.350, 0 ms, IPDFExportFeatures::get_MetaDataWritingParams( (IDocumentContentInfoWritingParams*__ptr64*)0xaff638 )
50800 :16:23:23.350, 0 ms, IDocumentContentInfoWritingParams::put_WriteCreationDate( WD_No )
50800 :16:23:23.351, 0 ms, IDocumentContentInfoWritingParams::put_WriteModificationDate( WD_No )
50800 :16:23:23.351, 34879 ms, IFRDocument::Export( (BSTR)"D:\ocr-test\NUIX\OCR-Test\OCR cache directory\7b6\temp7b6d16de5d4d1c12f4c3ce4d220d4f87.pdf", FEF_PDF, (structIUnknown*)0x3af1c5f4 )

Timers can be improved by using the MPM_Parallel setting which will allow the Engine to work with several pages at a time using the maximum allowed capabilities of the machine and increasing recognition speed drastically. 

 

 

 

 

 

 

 

Have more questions? Submit a request

Comments

0 comments

Please sign in to leave a comment.