More efficient convolution algorithm with a diagonal matrix as a kernel (gauss filter) for memory and speed improvement to calculate the filtered correlation matrix for boundaries as zero-padded or periodic and temporal correlation.
More efficient convolution algorithm with a diagonal matrix as a kernel (gauss filter) for memory and speed improvement to calculate the filtered correlation matrix for boundaries as zero-padded or periodic and temporal correlation.