An Image Matching Recognition Approach Based on ICM Icon

摘要

The scale and rotation change of target make the effect of image matching recognition worse. In order to resolve this problem, a novel recognition algorithm based on intersecting cortical model (ICM) icon was introduced, which provides an invariant representation of the object's rotation, scaling and translation. First, the real-time map and the reference map were input to ICM model to get icons. Then, similarity is analyzed between two icons to realize object matching recognition. This algorithm not only works well when the target has the scale change or rotation, but also has the low operation cost. Last, the experiments show the new method's effectiveness and superiority.