The Mask Tuning Scheme for Texture Feature Extraction and Its Parallel Implementation

This paper describes the parallel extension of the mask tuning scheme for texture feature extraction. Unlike other parallel systems in which specific parallel computer architectures are required, our parallel solution to increase processing speed is examined by introducing the concept of remote procedure call(RPC) in distributed systems for Iocal or remote interprocess communication. Based on our previous sequential implementation of the dynamic texture feature extraction associated with a texture 'tuned' mask, the rotational and scale invariant texture classification system is consequently established in parallel by sharing and transferring information on local machines via the network.