Image analysis using coefficient distributions with selective basis feature representation

摘要

Distributional information for a set of α vectors is determined using a sparse basis selection approach to representing an input image or video. In some examples, this distributional information is used for a classification task.

主权项

1. A computer-implemented method for automated image feature processing comprising:
accepting a data representation of a plurality of m dimensional feature vectors xs representing a processing of an image or video signal; accessing a dictionary of N basis vectors, where N>m; for each feature vector xs, forming using a computer a representation of the feature vector using a selection of less than all of the basis vectors of the dictionary, the representation including coefficients αs,n corresponding to the selected basis vectors; for each dictionary basis vector n, determining using a computer distribution characteristics of the coefficients αs,n over the plurality of feature vectors; combining the distribution characteristics corresponding to the plurality of basis vectors to form a combined feature vector; applying the combined feature vector to a computer-implemented classifier to determine a classification of the image or video signal.