A New Multi-scale Gaussian Interpolator that Models the Blind Spot in Human Eye

The "filling-in" mechanism of the blind spot in human retina has been modeled based on a possible response function for the non-classical receptive field of retinal ganglion cells. The same function, it has been shown previously, can also be used for the construction of a highly informative "raw primal sketch" in early vision. This substantiates the claims made from psychophysical experiments that the "filling-in" mechanism of blind-spot and edge detection are more or less co-occurring events in vision. The new model finds important application in providing a powerful vision tool for performing "near-ideal" interpolation with a linear combination of four zero-mean multi-scale Gaussian functions