A multicore architecture focused on accelerating computer vision computations

Abstract

This paper deals with accelerating computer vision computations using a specialized multicore architecture. Computer vision is one of the fastest-evolving segments of computer science. Even though computer vision uses time-consuming methods, the processing can be accelerated using specialized multicore processor architectures. Single-core processors are a legacy, since they have reached their physical limits. The way to go is to use multicore architectures, which can be also used to accelerate computations in specialized areas such as computer vision. This paper describes a specialized multicore architecture that can be used to accelerate time-consuming calculations in the field of computer vision. The architecture proposed in this paper belongs to the Harvard architecture family.