A two-dimensional nonuniform sampling expansion model (Q1314546)

We were not able to reproduce the numerical examples in the paper. The matrix of nonuniformly spaced samples \(g_{4 \times 4}\) given in Table 2 on Page 8 is not correct according to the definition given in the text. For example, the first element is given as 7.692. But according to the definition of \(g(u,v)\), the sample value at the top-left location of \(g_{4 \times 4}\) should be: \[ \begin{multlined} g(0.152, 0.152) = 50 \biggl[ 0.5 + \cos \bigl \{2 \pi (0.152 \times 0.05 + 0.152 \times 0.05) \bigr\} + \cdots \biggr] = \\ 50[1.499 + \cdots]\text{ above 200}. \end{multlined} \] Our computation does not yield the same numerical value as in the paper. Accordingly, the following 2D-DCT values also do not agree with the ones in the paper. We used our own nonuniform sample values for reconstruction. The reconstruction seems reasonable but can not be related to any results in the paper in an obvious way.