An Algorithm of Inverse Log Polar Transform Based on Neighbor Pixels

Log polar transform is not only a mathematical description of the retino-cortical mapping of human vision system, but also a vital algorithm of space-variant resolution theory. The obverse transform algorithm leads to pixel discreteness in the fovea area. Aiming at solving this problem, the inverse log polar transform algorithm based on grey weight value of neighbor pixels is presented. According to this algorithm, the acquired mapping transform image is continuous, meanwhile the quality of restored image is improved efficiently. Finally a bilinear interpolation method is adopted to decrease the computing burden of the algorithm.