Abstract: The present invention provides a technique for model improvement in supervised learning with potential applications to a variety of imaging tasks, such as segmentation, registration, detection. In particular, it has shown potential in medical imaging enhancement.

Abstract: A method of training a prediction tool to generate at least one synthetic full-contrast image from zero-contrast and low-contrast images of a subject may involve receiving a training set a set of images of a set of subjects, the images of each subject comprising a full-contrast image, a low-contrast image, a first zero-contrast image acquired prior to the acquisition of the full-contrast image, and a second zero-contrast image acquired prior to the acquisition of the low-contrast image. An artificial neural network may be trained with the training set by applying the first and second zero-contrast images from the set of images and the low-contrast images from the set of images as input to the artificial neural network and using a cost function to compare the output of the artificial neural network with the full-contrast images from the set of images to train parameters of the artificial neural network using backpropagation.

Abstract: The present invention relates to the technical field of machine learning. Subject matter of the present invention is a novel approach for training a neural network and the use of this approach for the processing of (medical) images.