Limitations of Connector Package
Upside-down image rotation:
We identify the top and bottom from the image as preprocessing of OCR. We use information such as the direction of arrangement of letters, letter / picture / chart layout, ruled line position, etc. to infer which is the correct rotation. Sometimes the top-bottom categorisation fails, for example in documents with a dark background / background such as photos.
Unusual fonts cannot be recognized:
In OCR, it is based on standardly used fonts such as Arial / Verdana / Times New Roman for foreign languages, MS Ming Dynasty / MS Gothic / MS UI Gothic for Japanese. Recognition may fail when using a special font, using extremely large / small font size.
Input image is not clear:
Although 200 dpi to 600 dpi is recommended, if the recognition rate is bad at 200 dpi, raise the resolution and scan again. If the input images are blurred or there is not enough contrast, the recognition rate will not improve. Please use clear images.
PDF output size is different:
Due to the limitation of the current engine, the size of the output PDF page may increase by about 0.3 to 1.7 mm.
PDF OCR in Japanese sometimes is impossible to print.
Word output layout collapse:
The OCR engine writes the character string of the recognition result as Word text and uses the font size / font type / line spacing / text width etc. to try to make it as close to the original as possible, but it is difficult to match with the original perfectly. Therefore, the following phenomena may occur: Although the manuscript was 2 pages, it became 3 pages / 1 page, margins increase or decrease
The footer's page number is incorrect in Word:
When outputting a document with a page number in the footer section as an input, Word automatically counts the page count and may be out of alignment with the page number of the original document. Example: Manuscript: 2 pages of 3/8, 4/8 Word output result: 2 pages of 1/2, 2/2 (Since there are only 2 pages, Word automatically judges 2 pages, 8 replaces 2 and redoes the page)
Doc format, that is created by OCR, cannot be displayed in iOS UIWebView:
It is said that iOS does not officially support doc / xls etc., it may not be displayed normally on iOS devices. This is because the OCR function creates data in RTF format and then outputs it to doc format. Please use docx / xlsx.
Excel output:
For Excel output, the colour of characters and cells is sometimes not reflected. Also, figures and photographs are not reflected. With regard to the part recognized as table, OCR engine uses the cell-ruled line / cell combination to create a table.
eFax:
The maximum possible number of fax pages is 40-50 A4 pages.
You cannot fill in To and Company fields in the fax cover page using this app.
Title is limited to 40 characters.
Body is limited to 255 characters.
When sending faxes abroad, eFax does not need an international access code '+‘.