Measuring the Similarity in 2D: the Case Study of the Extremely Low Frequency Magnetic Field

This paper introduces a new patch-based similarity measure for 2D arrays which is computed by considering not only the value but also the position of the elements in the bi-dimensional arrays. The similarity is evaluated as the mean area of the largest square patches which are in common at the same position in the two arrays. The proposed approach is tested in a case study where considering the position of the patches in the bi-dimensional space is very important. Specifically, it is used as a performance measure in predicting the magnetic field levels generated by a laptop computer and represented as a danger map in the bi-dimensional space. The obtained results are very promising.