OcrXmlOutputOptions Enumeration

The various IOcrDocument.SaveXml methods accept a combination of one or more of the OcrXmlOutputOptions enumeration members to control the format of the output XML data.

The format of the result XML data is as follows:

<?xml version="1.0" encoding="UTF-16" standalone="yes"?> <pages> <page> <zone> <paragraph> <line> <word> <character/> <character/> </word> </line> </paragraph> </zone> </page> </pages>

The pages element is repeated once per document and it has no value and no additional attributes.

The page element is repeated for every page in the document (IOcrDocument.Pages.Count). If this page has not been recognized or contains no zones, then the page element will not contain any child zone elements.

The page element has no value and contains the following additional attributes:

The zone element is repeated for every zone in the current page (IOcrPage.Zones). The zone element has no value and contains the following additional attributes:

The paragraph element is repeated for every text paragraph in the current zone and it has no attributes. If this zone has no recognition text, then the paragraph element will not contain any child line elements.

The paragraph element has no attributes and no value.

The line element is repeated for every line of text in the current paragraph. The line element has no value and contains the following additional attributes:

The word element is repeated for every word of text in the current line. If OcrXmlOutputOptions.Characters was not specified in the generation options; then the word element will contain the value of the word as its value. Otherwise, the word element will contain no value.

The word element has the following attributes:

The character element is repeated for every character in the following word only if OcrXmlOutputOptions.Characters was specified in the generation options. Otherwise, the word element will contain no child character elements. If OcrXmlOutputOptions.Characters was specified in the generation options; then the character element will contain the value of the character as its value. Otherwise, the character element will contain no value.

The character element contains the following additional attributes:

The following is an example of the XML output when OcrXmlOutputOptions.None is specified:

<?xml version="1.0" encoding="UTF-16" standalone="yes"?> <pages> <page horizontal_resolution="300" vertical_resolution="300" width="2544" height="3294"> <zone type="Text" left="371" top="370" right="831" bottom="420" subtype="Text" recognition_module="Auto" fill_method="Default"> <paragraph> <line left="372" top="371" right="830" bottom="419" base="29"> <word left="372" top="371" right="554" bottom="409" base="30">License</word> <word left="570" top="372" right="830" bottom="419" base="29">Agreement</word> </line> </paragraph> </zone> </page> </pages>

Here is the same XML output when OcrXmlOutputOptions.Characters is specified:

<?xml version="1.0" encoding="UTF-16" standalone="yes"?> <pages> <page horizontal_resolution="300" vertical_resolution="300" width="2544" height="3294"> <zone type="Text" left="371" top="370" right="831" bottom="420" subtype="Text" recognition_module="Auto" fill_method="Default"> <paragraph> <line left="372" top="371" right="830" bottom="419" base="29"> <word left="372" top="371" right="554" bottom="409" base="30"> <character left="372" top="372" right="398" bottom="408" base="36" confidence="100">L</character> <character left="402" top="371" right="409" bottom="408" base="37" confidence="100">i</character> <character left="414" top="381" right="438" bottom="409" base="27" confidence="100">c</character> <character left="442" top="381" right="468" bottom="409" base="27" confidence="100">e</character> <character left="472" top="381" right="496" bottom="408" base="27" confidence="100">n</character> <character left="501" top="381" right="525" bottom="408" base="27" confidence="100">s</character> <character left="529" top="381" right="554" bottom="408" base="27" confidence="100">e</character> </word> <word left="570" top="372" right="830" bottom="419" base="29"> <character left="570" top="372" right="604" bottom="408" base="36" confidence="100">A</character> <character left="607" top="381" right="633" bottom="419" base="27" confidence="100">g</character> <character left="639" top="381" right="655" bottom="408" base="27" confidence="100">r</character> <character left="657" top="381" right="682" bottom="408" base="27" confidence="100">e</character> <character left="685" top="381" right="710" bottom="408" base="27" confidence="100">e</character> <character left="715" top="381" right="753" bottom="408" base="27" confidence="100">m</character> <character left="758" top="381" right="783" bottom="408" base="27" confidence="100">e</character> <character left="788" top="381" right="812" bottom="408" base="27" confidence="100">n</character> <character left="815" top="374" right="830" bottom="408" base="34" confidence="100">t</character> </word> </line> </paragraph> </zone> </page> </pages>

You can use the OcrXmlOutputOptions.CharacterAttributes option along with IOcrDocumentManager.GetFontNames to obtain the font family name of each character. When performing OCR, the engine cannot distinguish similar fonts such as Arial and Calibri, instead, the engine gets information on whether the character has serif and whether the font is proportional or fixed.

According to GetFontNames, the returned value is an array of fonts as follows:

We are only interested in the first four values and with this information. As described above, using the CharacterAttributes option will add the following attributes to the element of each character in the XML: "proportional, serif, bold, italic". With this information and the font names obtained from GetFontNames, you can easily conclude the font family name. The following code snippet accomplishes this:

             private static void ShowFontNames(IOcrDocumentManager ocrDocumentManager, string xmlFileName)
             {
                // Get the Latin font names
                string[] fontNames = ocrDocumentManager.GetFontNames("en");
            
                // Note, you can also use a fixed array for most cases. The font names are not changed by the OCR engines
                // unless changed manually by the user:
                /*
                string[] fontNames =
                {
                   "Times New Roman",   // Proportional Serif
                   "Arial",             // Proportional Sans-Serif
                   "Courier New",       // Monospace Serif
                   "Arial"              // Monospace Sans-Serif
                };
                */
            
                // Read the XML file
                XPathDocument xmlDoc = new XPathDocument(xmlFileName);
                XPathNavigator nav = xmlDoc.CreateNavigator();
                XPathNodeIterator charactersIter = nav.Select(@"//character");
                while(charactersIter.MoveNext())
                {
                   XPathNavigator characterNav = charactersIter.Current;
            
                   // Read its attributes
                   bool isSerif;
                   bool isProportional;
            
                   if(string.Compare(characterNav.GetAttribute("serif", characterNav.NamespaceURI), "yes", StringComparison.InvariantCultureIgnoreCase) == 0)
                      isSerif = true;
                   else
                      isSerif = false;
            
                   if(string.Compare(characterNav.GetAttribute("proportional", characterNav.NamespaceURI), "yes", StringComparison.InvariantCultureIgnoreCase) == 0)
                      isProportional = true;
                   else
                      isProportional = false;
            
                   // Now find the font based on this info
                   int index;
            
                   if(isProportional && isSerif)
                      index = 0;
                   else if(isProportional && !isSerif)
                      index = 1;
                   else if(!isProportional && isSerif)
                      index = 2;
                   else
                      index = 3;
            
                   string fontName = fontNames[index];
                   Console.WriteLine("Character: {0} font: {1}", characterNav.Value, fontName);
                }
             }
             

System.Object
   System.ValueType
      System.Enum
         Leadtools.Forms.Ocr.OcrXmlOutputOptions

Target Platforms

Reference

Leadtools.Forms.Ocr Namespace
Leadtools.Forms.DocumentWriters.DocumentFormat
IOcrDocumentManager Interface
IOcrDocument Interface
IOcrDocument.Save
IOcrDocument.SaveXml
IOcrPage.Recognize
IOcrEngine Interface
OcrEngineManager Class
OcrEngineType Enumeration
Programming with the LEADTOOLS .NET OCR
Files to be Included with Your Application