There is too much noise. I reprocessed the image using scantailor: scantailor-cli -v --layout=1.5 --dpi=300 --output-dpi=300 --margins-top=5 --margins-bottom=5 --margins-left=10 --margins-right=10 --threshold=-40 --picture-shape=rectangular --enable-page-detection --white-margins --enable-fine-tuning --color-mode=black_and_white --normalize-illumination=true --despeckle=off --start-filter=3 0002_nrm.png out No panick on this long command: it is my standard scantailor one. With no picture I use '--color-mode=black_and_white'. With a picture into the image I use '--color=color_grayscale' followed by an imagemagic "convert $f -type GrayScale -normalize $f" Note the thershold=-40 (negative 40) , this is due to the thick ink in your image. Usually I do -threshold=15 (positive 15).
<https://lh3.googleusercontent.com/-l7A_6z7YJlU/Ux7nD-Sh2YI/AAAAAAAAAC0/CfWB_mjxPDs/s1600/0002_nrm.tif> Got this translation : /opt/out> ts 0002_nrm.tif a40 Tesseract Open Source OCR Engine v3.03 with Leptonica /opt/out> cat a40.txt POPULAR TALES 0' THE WEST HIGHLANDS. + XVIII. T H E C H E S T. >From Mrs MacGeechy. Inlay. EFORE this there was a king, and he wished to see his son with a wife before he should depart. His son said he had better go for a wife ; and he gave him half a hundred pounds to get her. He went for- ward the length of a day, and when the night came he went in to a hostelry to stay in it. He went down to a chamber with a good fire in front of him; and when he had gotten meet. the man of the house went down to talk to him. He told the man of the house the jour- ney on which he was The man of the house told him he need not go further; that there was a little house opposite to his sleeping chamber; that the man of the house had three fine daughters ; and if he would stand in the window of his chamber in the morning, that he would see one after mother coming \0 dress herself. That they were all like each other, and that he could not one from the other. but that v01. IL n - \ Note that in the word 'fire' and 'fine' there is a sign 'fi' whch is in fact a single character and is a pleague in the unicharset I added this rule in my FRA module in unicharambigs to get rid of it: 1 fi 2 f i 1 1 fl 2 f l 1 You will note that I put 1 at last --> mendatory replace. Who introduced these two signs and why, I don't know but they are extremly domageable. -- -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/tesseract-ocr?hl=en --- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
