| CPC G06N 3/045 (2023.01) [G06V 10/761 (2022.01); G06V 10/82 (2022.01)] | 7 Claims |

|
1. A mental image visualization method comprising:
generating, by a deep neural network (DNN) trained using a data set of feature training images for feature learning, sample images in which different objects in a category same as a category of objects captured in the feature training images are captured;
inputting the sample images to the DNN;
obtaining feature vectors of the sample images from the DNN, the feature vectors each resulting from one of the sample images being converted by the DNN into an n-dimensional vector, where n is an integer greater than or equal to 100, and the feature vectors are used to generate an image showing a mental image;
obtaining eigenvalues by applying dynamic mode decomposition (DMD) to feature vectors obtained by weighting the feature vectors of the sample images, according to results of sensory evaluation on the sample images by psychological reverse correlation;
selecting at least two eigenvalues from among the eigenvalues and obtaining at least two eigenvectors each having one of the at least two eigenvalues; and
generating, using the DNN, at least two images showing at least two sub-mental images from the at least two eigenvectors, the at least two sub-mental images forming the mental image and being not on condition of mutual orthogonality.
|