PSYCH 221 FINAL PROJECT

L1 & L2 - Norm
    L1 & L2 Norm:

    L2 Norm:

    L2 Norm measures the Euclidian distance between the pixels of the original and the corrected image. Formula for the L2 norm is:

    ∑((img_original _i-img_corrected_i)^2),

    where the summation index i goes over all the pixels in the image. A noteworthy feature of L2 Norm is that the contribution of a specific pixel grows quadratically as its distance to the corresponding pixel grows. Hence, L2 norm “punishes” much more for bigger distances than L1 norm. 

    L1 Norm:

    Norm measures the absolute distance between the pixels of the original and the corrected image. Formula for the L1 norm is:

    ∑(abs(img_original _i-img_corrected_i)),

    where the summation index i goes over all the pixels in the image. Unlike in L2 norm, the contribution of a distance between two pixels grows linearly with the distance. So, L1 norm “punishes” relatively more for smaller distances than L2 norm.