ABSTRACT
Energy minimization plays a central role in computer vision algorithms. Nanomagnets offer a tantalizing alternative to traditional form of digital computing for solving quadratic energy minimization problems, drastically reducing the computational time required. This chapter demonstrates the viability of using single domain nanomagnetic coupling for function minimization computing. There have been proposals for the use of nanomagnets to directly solve quadratic minimization problems, especially those arising in computer vision applications. Imagine an energy minimization co-processor based on nanomagnets that are heterogeneously integrated with complementary metal–oxide–semiconductor. There have been proposals for using regular arrays of quantum dots and nanomagnets for low-level vision, mainly segmentation where the input and the output are both regular grid of pixels. The basic unit of computation for magnetic logic is a nanomagnet with dimensions and materials such that it exhibits single domain behavior, that is, it can be modeled as one overall magnetic state.
