Bitmap images recognition by dynamical neural network defined in the space of complex numbers

A novel method of bitmap images recognition is considered for the case of introducing two parameters per each image pixel. The image is being mapped to the vector with complex elements corresponding to the pixel characteristics of the original image. A model of the dynamical competitive neural network performs the recognition where each artificial neuron is defined by the complex activity level. Only the activity level of the neuron corresponding to the most matching pattern to the offered one achieves nonzero value at the end of the recognition process.