Image processing method, image processing apparatus, program, storage medium, production apparatus, and method of producing assembly
AssigneeCanon Kabushiki Kaisha
InventorHiroshi Kitajima
Life of this patent
Estimating the term…
Abstract
A tentative local score between a point in a feature image in a template image and a point, in a target object image, at a position corresponding to the point in the feature image is calculated, and a determination is performed as to whether the tentative local score is smaller than 0. In a case where the tentative local score is greater than or equal to 0, the tentative local score is employed as a local score. In a case where the tentative local score is smaller than 0, the tentative local score is multiplied by a coefficient and the result is employed as a degree of local similarity.
Claims27
An image processing method for searching for a position of a pattern included in a model image in a search target by using a plurality of template images and a pyramid, the plurality of template images being hierarchized sequentially by performing reduction processing and low resolution processing on a feature image of the model image, the pyramid being obtained by search target image hierarchization in accordance with reduction and resolution of the template images, the method comprising:
- calculation of a degree of similarity being performed between the plurality of template images and the search target image, the calculation being performed sequentially in an ascending order of resolution from low to high, the calculation being performed through procedure comprising:
- a local degree of similarity calculation process of calculating local degree of similarity between a point of the feature image of the model image and a point corresponding to the point of the feature image in the search target image for each range corresponding to the feature image of the model image while shifting a position of the feature image of the model image on the search target image;
- and
- a degree of similarity calculation process of calculating degree of similarity for each range corresponding to the feature image of the image model at each position of said each local degree of similarity,
- wherein correction is performed to decrease an absolute value of the local degree of similarity in a case where the local degree of similarity is less than a predetermined value in the local degree of similarity calculation process.
An image processing method for performing image processing by an image processing apparatus using a first pyramid including a plurality of template images having different resolutions and hierarchized in layers according to the resolutions, a second pyramid including a plurality of target object image having different resolutions from each other but equal to the respective resolutions of the template images in the first pyramid and hierarchized in layers according to the resolutions such that an image similar to a feature image included in a template image in the first pyramid is searched for from a target object image in the second pyramid by evaluating similarity between images in corresponding layers on a layer-to-layer basis in the order of resolution from the lowest to highest,
- the feature image being a set of edge points extracted via an edge extraction process,
- the target object image being an image including a set of edge points extracted via the edge extraction process,
- an edge strength at each edge point being represented by a vector,
- the method comprising:
- calculating a degree of local similarity between a point in the feature image and a corresponding point in the target object on a point-by-point basis for each of all points in the feature image;
- and
- calculating the degree of the similarity between the feature image and the target object image by determining the sum of the calculated degrees of local similarity and normalizing the sum,
- the calculating the degree of local similarity including
- calculating the inner product between a first vector representing the edge strength at an edge point in the feature image and a second vector representing the edge strength at a point, in the target object image, at a position corresponding to the edge point in the feature image, and
- in a case where inner product is greater than or equal to a predetermined value, calculating the degree of local similarity by dividing the inner product by the product of the magnitude of the first vector and the magnitude of the second vector, while in a case where inner product is smaller than the predetermined value, setting the degree of local similarity to be equal to a value whose absolute value is smaller than the value obtained by dividing the inner product by the product of the magnitude of the first vector and the magnitude of the second vector.
An image processing apparatus comprising:
- an image acquiring unit configured to acquire a captured image;
- and
- a processor configured to search for a position of a pattern included in a model image the acquired captured image
- wherein the processor searches the position of the pattern using a plurality of template images and a pyramid, the plurality of template images being hierarchized sequentially by performing reduction processing and low resolution processing on a feature image of the model image, the pyramid being obtained hierarchization of the captured image in accordance with reduction and resolution of the template images by calculation of a degree of similarity being performed by the processor between the plurality of template images and a search target image generated from the captured image, the calculation being performed sequentially in an ascending order of resolution from low to high, the calculation being performed through procedure comprising:
- generating the search target image via an edge extraction process of the captured image;
- calculating local degree of similarity between a point of the feature image of the model image and a point corresponding to the point of the feature image in the search target image for each range corresponding to the feature image of the model image while shifting a position of the feature image of the model image on the search target image;
- and
- calculating degree of similarity for each range corresponding the feature image of the model image at each position on the basis of said local degree of similarity;
- wherein correction is performed to decrease an absolute value of the local degree of similarity in a case where the local degree of similarity is less than a predetermined value in the local degree of similarity calculation process.
Dependencies are inferred from the claim text.