Method for the computer-assisted learning of an artificial neural network for detecting structural features of objects

A method for computer-aided training of an artificial neural network for recognizing structural features on plants or on plant constituents for cloning the plants, whereby the structural features represent cutting positions on the plants, wherein the network used is a convolutional neural network (CNN) having a multiplicity of neurons, said network comprising a multiplicity of convolutional and/or pooling layers for extraction of information from images of the plants or the plant constituents having structural features to be recognized for a classification of the features by further layers, comprising:

• scaling the images up and down from layer to layer;
• during the scaling of the images from layer to layer obtaining information about the structural features of the plants or the plant constituents, specifically independently of the scaling of the images;
• and
• labeling or designation of the structural features of the plants or the plant constituents semiautomatically on the images in preparation for a training process of the neural network, wherein the features and/or the plants or the plant constituents are isolated from a background of the images and labelling of the features is carried out by a person,
• wherein a plurality of images from different perspectives of a same one of the plants or the plant constituents for recognizing the structural features are transferred simultaneously to the neural network,
• wherein computer-aided operations for recognizing the structural features of the plants or the plant constituents are carried out on the images in parallel on a plurality of GPUs,
• wherein the plurality of images of a same one of the plants or the plant constituents are scaled or rescaled and stitched prior to transfer to the neural network,
• wherein the labeling of the structural features of the plants or the plant constituents in a computer-aided manner is proposed.