Data compression using non - Uniform sampling

This paper is concerned with the problem of non-uniform sampling and reconstruction of data. A fast and robust method for approximation of contour from a set of non-uniformly distributed sample points is described here. Data is represented by irregular samples distributed in space. The samples are chosen along the contour with spacing determined by the error between the original contour and the contour reconstructed by uniform set of sample points. Lagrange's interpolation is used to reconstruct the contour back from non-uniform set of samples. The algorithm iteratively finds the minimum number of samples along the contour. The error analysis between original and reconstructed contour reveals that iterative algorithm of selecting irregular samples is satisfactory.