Which Imatest charts support automatic detection?

Open

The best experience with Imatest comes when it can automatically detect targets. This allows for tests to be automated and improves the efficiency and repeatability of your testing workflow (see following sections).

Targets with special autodetection

• SFRplus—requires bars on top and bottom of chart to be present in the image

• 24-patch X-Rite Colorchecker—Works best without extreme distortion

• Checkerboard—We can detect square areas of checkerboards that are not excessively bloomed or blurred

• Rezchecker—We can detect the version of this target with 4 dots

• Dot pattern 

• Concentric Rings FOV—We detect a single registration mark to find the chart center

Targets detected with four or more registration marks

• 36-Patch Dynamic Range Test Charts

• Arbitrary Charts—Supports SFRreg-style registration marks or black circles

• eSFR ISO 12233:2014+

• ISO-14524

• ISO-15739

• Random Patern

• SFRreg

• Sinusoidal Siemens Star

• Spilled coins: Black & White, Color, Cross-correlation

• Transmissive Color Target

Imatest plans to implement Colorchecker SG autodetection in a future release. Note that we will not perform autodetection for the Obsolete ISO 12233:2000 target because it can easily produce invalid results. 

Registration mark options for automatic detection

Starting in Imatest version 24.1, users can choose between the default registration mark detection and an alternative machine learning/computer vision (ML/CV) registration mark detection with a tunable confidence threshold. This new setting can be accessed from the Imatest Master main window choosing Settings, General Settings, then navigating to the settings for a particular target via the analysis groupings.

Starting in Imatest version 24.2, additional registration mark detection settings are available.

• Users can choose to apply a High Noise correction for registration mark detection. This may help with failing regmark detection for scenes with poor SNR.

• Users can choose to Skip Refinement Step for registration mark detection. This may help with failing regmark detection at the cost of less accurate localization.

Open