Corner Detection

Can be one of the following values:

The 'Method' property selects the algorithm for corner detection:

• 'Square' detects non-skewed 90° corners, ideal for forms and boxes.
• 'Harris' detects corners at arbitrary angles, suitable for general layouts.

Tuning Guidance:

• Use 'Square' for structured forms; use 'Harris' for natural or rotated content.
• Review diagnostic overlays to compare results.

The 'Max Corners' property limits the number of detected corners to avoid excessive or redundant results.

• Increase to detect more corners in complex layouts.
• Decrease to focus on the most prominent corners.

Tuning Guidance:

• Set based on the expected number of corners in your documents.
• Review the "Corners" diagnostic image to ensure all relevant points are detected.

Can be one of the following types:

The 'Binarization' property allows you to preprocess the image using configurable binarization settings. This can improve corner detection on noisy, low-contrast, or color images.

• Enable and configure for challenging images.
• Leave disabled for clean, high-contrast scans.

Tuning Guidance:

• Review the "Binarized" diagnostic image to verify preprocessing.
• Adjust binarization parameters as needed for your document set.

The 'Corner Threshold' property sets the minimum response value for a point to be considered a corner. Lower values detect more corners (including weak ones); higher values detect only strong corners.

• Decrease to increase sensitivity.
• Increase to reduce noise and false positives.

Tuning Guidance:

• Review the "Corners" and "Harris Response" diagnostics to see the effect.
• Adjust to balance between missing corners and detecting noise.

Can be one of the following values:

The 'Filter Type' property selects the derivative operator for the Harris method.

• 'Sobel' is the most common and provides good edge localization.

Tuning Guidance:

• Use the default unless you have specific requirements.
• Only applies when 'Method' is set to 'Harris'.

Can be one of the following values:

The 'Filter Size' property sets the size of the convolution kernel for the Harris method.

• 3x3 is standard for most documents.
• Larger sizes provide more smoothing but may miss fine details.

Tuning Guidance:

• Use 3x3 for high-resolution or clean images.
• Only applies when 'Method' is 'Harris' and 'Filter Type' is 'Sobel'.

The 'Window Size' property sets the size of the neighborhood considered for each corner candidate in the Harris method. Larger windows smooth out local variations but may miss fine details.

• Increase for noisy images.
• Decrease for fine detail.

Tuning Guidance:

• Adjust based on the scale of features in your documents.
• Only applies when 'Method' is 'Harris'.

The 'K Parameter' controls the sensitivity of the Harris detector to corners versus edges. Lower values detect more corners; higher values detect fewer.

• Typical values range from 0.04 to 0.06.

Tuning Guidance:

• Start with the default and adjust for your document type.
• Only applies when 'Method' is 'Harris'.

The 'Mask Size' property sets the size of the neighborhood for the Square method. Larger sizes detect larger corners; smaller sizes detect finer details.

• Adjust to match the scale of corners in your forms.

Tuning Guidance:

• Only applies when 'Method' is 'Square'.

The 'Border Size' property sets the width of the border region used in the Square method. Adjust to control how close to the edge corners are detected.

• Increase to avoid detecting corners near the page edge.
• Decrease to include edge corners.

Tuning Guidance:

• Only applies when 'Method' is 'Square'.

The 'Min Confidence' property sets the minimum response value for a point to be accepted as a corner in the Square method. Lower values detect more corners; higher values detect only strong corners.

• Adjust to balance sensitivity and noise.

Tuning Guidance:

• Only applies when 'Method' is 'Square'.

The 'Command Type Name' property shows the user-friendly name of the IP Command as it appears in the Grooper UI. This helps users identify which command is being configured or executed, especially when working with IP Profiles containing multiple commands.

This property is read-only and is automatically set based on the command's type.

The 'Supported Pixel Formats' property lists the image pixel formats that the IP Command can process. If an image does not match one of these formats, the command will be skipped for that image.

• By default, all pixel formats are supported unless a derived command restricts this list.
• Example formats include Pixel8bppGrayscale, Pixel24bppBgr, and Pixel1bppIndexed.

Use this property to verify compatibility when building IP Profiles or troubleshooting why a command did not execute.