Read Text (OCRMax) Identification Tool - Segmentation Tab
The Segmentation tab is used to adjust and modify the settings that segment characters in the region. Typically, the default values and settings will read most text, however, in more challenging cases, some parameter values may need to be adjusted. In these types of cases, specifying a minimum character width, a maximum width and/or a minimum pitch (i.e. the character-to-character distance; for example, from the left edge of one character to the left edge of the one after it) will typically address the issues. For difficult cases, and to help speed the creation of the font database, use the Auto-Tune button to launch the Auto-Tune dialog and allow the tool to automatically calculate the optimal segmentation settings.
Segmentation Guidelines
- Images that contain backgrounds with strong textures and/or so much noise that the characters blend into the background will be difficult to segment.
- The region should only contain the characters to be read, and not any extraneous strong features, such as other characters or label edges.
- In instances where two characters touch, most likely, parameters will need to be adjusted. Fixed-width touching characters can be compensated for by specifying the character width; however, proportional fonts with touching characters are problematic, with the function correctly processing some cases, but not others.
- In cases involving short lines of text (for example, three or fewer characters) or relatively short lines with a lot of line jitter, specifying the expected Angle Range should help the function properly compensate for the inherent uncertainty in determining the orientation of a short line of text.
- All the characters in a line of characters must have the same orientation and skew.
- For well-separated dot-matrix print (e.g. where the dots are not touching), adjusting parameters such as the Minimum Character Fragment Size may be necessary to properly segment the characters.
- Character stroke width must be greater than or equal to two pixels.
- The minimum character size for large characters (typically alphanumeric) is 8 x 8 pixels. The minimum character size for small characters (such as periods) is 2 x 2 pixels.
Auto-Tune
Launches the Auto-Tune dialog, which is used in conjunction with the Read Text (OCRMax) Identification Tool to configure the segmentation parameters of the tool, and train a font database, through the construction of Auto-Tune Records to create an Auto-Tune Model. An Auto-Tune Record consists of an image, the training string and the user-verified location of the characters within the image. The Auto-Tune Model consists of one or more Auto-Tune Records, plus the settings that best segments all of the images within the Auto-Tune Records.
- Cognex recommends that when using the Auto-Tune dialog, the segmentation tuning be performed on an In-Sight Emulator instead of an In-Sight vision system. The Read Text (OCRMax) tool's Auto-Tune algorithm requires a large amount of computational resources, and performance will be better on the Emulator versus a vision system.
- When training a font database and using the Auto-Tune dialog to segment the characters, Cognex recommends beginning the process with images that best represent how the characters should look (i.e. use your best images first), working from best to worst. This will allow the Read Text (OCRMax) tool's Auto-Tune algorithm to tune more successfully.
Character Polarity
Defines the polarity of the characters in the input image: Black on White, White on Black or Auto (default).
- Black on White: The polarity of the text is black text on a white background.
- White on Black: The polarity of the text is white text on a black background.
- Auto: The tool will automatically determine the polarity of the text and background.
Character Width Type
Defines how the widths of the characters in the font are expected to vary: Auto (default), Fixed or Variable.
- Auto: The character width is unknown; the font may have either fixed or proportional width.
- Fixed: All of the character rectangles in the font have the same width.
- Variable: The characters in the font may have character rectangles with different widths.
Minimum Character Width
Defines the minimum width of a character's character rectangle, in pixels (1 - 1000; default = 3), that a character must have to be reported.
Minimum Character Height
Defines the minimum height of a character's character rectangle, in pixels (1 - 1000; default = 3), that a character must have to be reported.
Use Maximum Character Width
Defines whether or not the tool should account for the maximum allowable width of a character's character rectangle (by default, the parameter is disabled).
Maximum Character Width
Defines the maximum width of a character's character rectangle, in pixels (1 - 5000; default = 100), that a character must have to be reported.
Use Maximum Character Height
Defines whether or not the tool should account for the maximum allowable height of a character's character rectangle (by default, the parameter is disabled).
Maximum Character Height
Defines the maximum height of a character's character rectangle, in pixels (1 - 5000; default = 100), that a character must have to be reported.
Use Minimum Character Aspect Ratio
Defines whether or not the function will account for the minimum allowable aspect ratio of a character, where the aspect ratio is defined as the height of the entire line of characters, divided by the width of the character’s character rectangle (by default, this parameter is enabled).
Minimum Character Aspect Ratio
Defines the minimum allowable aspect ratio (0-500; default = 80) of a character.
Angle Range
Defines the angel search range (0 - 45; default = 0), in degrees.
Skew Range
Defines the skew search range (0 - 45; default = 0), in degrees.
Show Diagnostics
Defines whether or not to display graphical diagnostic data on the image. When enabled, Regions will be drawn around kept characters (green), rejected characters (red), kept fragments (cyan), rejected fragments (yellow), the mainline of the text (blue) and the extended mainline of the text (magenta).