Memristive pulse coupled neural network with applications in medical image processing

Medical imaging has become an integral part of modern medical technology increasingly. The image information provided by multimode imaging technology can complement each other. Fusing computed tomography (CT) and magnetic resonance imaging (MRI) images can combine their unique information to diagnose brain diseases by radiation therapy. And medical imaging process is easy to produce noise, which will impact doctors on diagnosis. So the medical image de-noising has important significance. In addition, the image edge extraction is helpful to the clinical diagnosis. So this paper constructs a tnemristive pulse coupled neural network (M-PCNN) for medical image processing. The memristance of Gale memristor decays exponentially with time, which can be used to adjust the threshold of pulse coupled neural network (PCNN) online. Integrating the memorability of memristor into PCNN makes the network have biological function. And the introduction of nanoscale memristor can also significantly reduce the scale of PCNN, which may promote the development of neural network hardware implementation. Numerical simulation verifies that the superiority of using this network in medical image fusion, image de-noising and image edge extraction.