Image processing method, image processing apparatus and non-transitory computer-readable storage medium therefor

摘要

An apparatus includes an extraction unit to extract a feature point in a first image. A determination unit determines a motion vector that corresponds to the feature point extracted from the first image by the extraction unit and a corresponding point that is located in a second image and that has a highest degree of similarity between image data for the feature point and image data for the corresponding point, among candidates of the corresponding point in the second image. A correction unit corrects a first motion vector determined by the determination unit by employing a second motion vector determined by the determination unit. The correction unit employs the second motion vector, in accordance with both of a first degree of similarity between the feature point and the corresponding point, and a second degree of similarity between the feature point and another candidate among the candidates.

主权项

1. An apparatus comprising:
an extraction unit configured to extract an edge feature point corresponding to an edge in a first image, from the first image; a first determination unit configured to determine temporally a temporal motion vector that corresponds to the edge feature point extracted from the first image by the extraction unit and a corresponding point that is located in a second image and that has a highest degree of similarity between image data for the edge feature point and image data for the corresponding point among candidates of the corresponding point in the second image; a setting unit configured to set a feature point, on a corner of the first image, regardless of whether the set feature point corresponds to an edge in the first image or not; a second determination unit configured to determine a motion vector corresponding to the edge feature point, by using a first temporal motion vector determined for the edge feature point by the first determination unit and a second temporal motion vector determined by the first determination unit, in accordance with the highest degree of similarity and to determine a motion vector corresponding to the set feature point set on the corner by the setting unit, using a temporal motion vector corresponding to a neighboring edge feature point and without using a temporal motion vector corresponding to the set feature point; and at least one processor configured to function as at least one of the extraction unit, the first determination unit, the setting unit, and the second determination unit.