FIELD: data processing.
SUBSTANCE: invention relates to mechanisms for detecting objects in images using neural networks. Method comprises obtaining an image representative of each element from a plurality of image elements using an input vector from a plurality of input vectors, each of the input vectors comprising one or more parameters associated with the visual representation of the corresponding image element; providing a plurality of input vectors to a first subnetwork of a neural network to obtain a plurality of output vectors, wherein each output vector of the plurality of output vectors is associated with an image element; determining, based on a plurality of output vectors, a subset of image elements relating to an image of an object, and determining location of image of object on image based on location of subset of elements.
EFFECT: technical result consists in providing automatic detection of objects on images.
20 cl, 10 dwg
