|
|
|
|
@ -25,7 +25,7 @@ do_binarization() {
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-IMG-BINPAGE
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-IMG-BIN
|
|
|
|
|
ocrd-olena-binarize -l $LOG_LEVEL \
|
|
|
|
|
-m mets.xml -I OCR-D-IMG -O OCR-D-IMG-BINPAGE \
|
|
|
|
|
-I OCR-D-IMG -O OCR-D-IMG-BINPAGE \
|
|
|
|
|
-p "$ocrd_olena_binarize_paramters"
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
@ -35,9 +35,9 @@ do_linesegmentation_tesserocr() {
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-SEG-REGION
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-SEG-LINE
|
|
|
|
|
ocrd-tesserocr-segment-region -l $LOG_LEVEL \
|
|
|
|
|
-m mets.xml -I OCR-D-IMG-BINPAGE -O OCR-D-SEG-REGION
|
|
|
|
|
-I OCR-D-IMG-BINPAGE -O OCR-D-SEG-REGION
|
|
|
|
|
ocrd-tesserocr-segment-line -l $LOG_LEVEL \
|
|
|
|
|
-m mets.xml -I OCR-D-SEG-REGION -O OCR-D-SEG-LINE
|
|
|
|
|
-I OCR-D-SEG-REGION -O OCR-D-SEG-LINE
|
|
|
|
|
|
|
|
|
|
# XXX compare ocrd-tesserocr-segment* vs tesseract native
|
|
|
|
|
}
|
|
|
|
|
@ -50,7 +50,7 @@ do_linesegmentation_sbb() {
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-SEG-REGION
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-SEG-LINE
|
|
|
|
|
ocrd-sbb-textline-detector -l $LOG_LEVEL \
|
|
|
|
|
-m mets.xml -I OCR-D-IMG-BINPAGE -O OCR-D-SEG-LINE \
|
|
|
|
|
-I OCR-D-IMG-BINPAGE -O OCR-D-SEG-LINE \
|
|
|
|
|
-p "$ocrd_sbb_textline_detector_parameters"
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
@ -60,7 +60,7 @@ do_ocr() {
|
|
|
|
|
ocrd_tesserocr_recognize_parameters='{ "model": "GT4HistOCR_2000000" }' # TODO mods:language + fontident → model
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-OCR-TESS
|
|
|
|
|
ocrd-tesserocr-recognize -l $LOG_LEVEL \
|
|
|
|
|
-m mets.xml -I OCR-D-SEG-LINE -O OCR-D-OCR-TESS \
|
|
|
|
|
-I OCR-D-SEG-LINE -O OCR-D-OCR-TESS \
|
|
|
|
|
-p "$ocrd_tesserocr_recognize_parameters"
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
@ -71,7 +71,7 @@ do_ocr_calamari() {
|
|
|
|
|
}'
|
|
|
|
|
ocrd workspace remove-group -rf OCR-D-OCR-CALAMARI
|
|
|
|
|
ocrd-calamari-recognize -l $LOG_LEVEL \
|
|
|
|
|
-m mets.xml -I OCR-D-SEG-LINE -O OCR-D-OCR-CALAMARI \
|
|
|
|
|
-I OCR-D-SEG-LINE -O OCR-D-OCR-CALAMARI \
|
|
|
|
|
-p "$ocrd_calamari_recognize_parameters"
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
@ -134,7 +134,7 @@ for ocr_filegrp in OCR-D-OCR-CALAMARI OCR-D-OCR-TESS; do
|
|
|
|
|
|
|
|
|
|
if ocrd workspace list-group | grep -q OCR-D-GT-PAGE; then
|
|
|
|
|
ocrd workspace remove-group -rf $ocr_filegrp-EVAL
|
|
|
|
|
ocrd-dinglehopper -m mets.xml -I OCR-D-GT-PAGE,$ocr_filegrp -O $ocr_filegrp-EVAL
|
|
|
|
|
ocrd-dinglehopper -I OCR-D-GT-PAGE,$ocr_filegrp -O $ocr_filegrp-EVAL
|
|
|
|
|
fi
|
|
|
|
|
|
|
|
|
|
done
|
|
|
|
|
|
