Recursive Gabor transform for images of very large size and design of massively parallel architecture

The conventional Gabor transforms, complete and generalized, cannot be applied for image matrices received as different partitions. In practice, very large images are received in the form of submatrix partitions. Then, taking transform on different partitions (of the same image) at different times is of prime importance. The situation can be easily handled if the transformation algorithm is made recursive. The authors present a recursive Gabor transform (RGT) amenable for pipelining partitions of different images interleaved. Apriori knowledge of image size is not required for the RGT unlike the conventional Gabor Transform. A dedicated massively parallel architecture based on Pacube VLSI arrays is presented which achieves a high degree of pipelining with respect to different image matrices. The pipelining rate is proportional to the delay of a Pacube macrocell in mode 1 operation