Learn More
In image processing, the images are represented as matrices of gray levels of pixels. In many operations performed on images, the new value of a pixel depends on its old value and values of its neighbours. One such operation is convolution. Its usage is described in 4]. One convolution can be eeciently computed in parallel as is shown in 1]. We will analyze(More)
Parallel programming is a good way for increasing the computational speed. Any network of workstations can be used as a relatively cheap parallel computer. We will show two methods for parallel computation of sequences of convolutions in this paper. They are pipelining and data partitioning. We will show the algorithms for these methods, performance(More)
  • 1