Classifying images

摘要

Methods, and apparatus for performing methods, for classifying an image. Methods include determining a corresponding set of metrics for each region of two or more regions of a pattern of regions of an image, and classifying the image in response to at least the corresponding set of metrics for each of the two or more regions of the pattern of regions.

主权项

1. A method of classifying an image, comprising:
determining a pattern of regions of an image by dividing the image in a predetermined grid, by using at least one of image-segmentation based tessellation, and detecting and selecting specific objects within the image, wherein the pattern of regions is determined based on historical data describing at least one of, portions of images that have been previously used to detect differences between authentic and counterfeit images, and portions of images where accurate copying is more difficult; for each region of two or more regions of the pattern of regions, determining a corresponding set of metrics; and classifying the image in response to at least the corresponding set of metrics for each of the two or more regions of the pattern of regions.