Decomposition and parallel architecture for the geometric transformation of digital images

摘要

The decomposition and implementation of image geometric transformation algorithms is considered. An approach based upon parallel within-row and column operations on image data is developed along with a taskspecific (modular) hardware for implementation. The time and hardware (space) complexity of this approach is compared with alternative architectures. Future research directions are indicated.