At the heart of any image recognition software is the ability to perform adaptive matching when comparing images for similarity. Just like in human recognition, adaptive matching enables recognition software to determine whether similar but not identical images are essentially the same.

The CogniSign Recognition Kernel software achieves superior results in adaptive matching because it finds the comparative transform between the images under evaluation as an integral part of the matching process. Competitor products rely on pre-processing techniques that use the comparative transform in a rudimentary form. This approach introduces matching errors and also reduces flexibility in application.