Image data processing method and apparatus
AssigneeHewlett-Packard Development Company, L.P.
InventorStephen B. Pollard
Life of this patent
Estimating the term…
Abstract
A method of processing image data generated using an image capture device comprising a lens, the method comprising, generating metric data for a plurality of regions of an image, the data representing a plurality of focus measures for each region generated at a plurality of lens positions, processing the metric data in order to cluster regions into at least one group on the basis of their respective focus measures, and determining a lens position for a group corresponding to a position of true focus for the group.
Claims21
A method of processing image data generated using an image capture device comprising a lens, the method comprising:
- generating, by a processor in the image capture device, metric data for a plurality of segments of an image, the metric data representing a plurality of focus measures for each of the segments generated at a plurality of lens positions;
- processing, by the processor, the metric data in order to cluster the segments into at least two groups on the basis of the respective focus measures of the segments, wherein each of at least one of the two groups includes multiple ones of the segments;
- and
- determining, by the processor, a lens position for each of the groups corresponding to a position of true focus for the corresponding group.
A method of processing image data generated using an image capture device comprising a lens, the method comprising:
- generating metric data for a plurality of regions of an image, the data representing a plurality of focus measures for each region generated at a plurality of lens positions;
- processing the metric data in order to cluster regions into at least one group on the basis of their respective focus measures;
- and
- determining a lens position for the group corresponding to a position of true focus for the group;
- wherein the focus measures are determined using a normalised power ratio metric generated for each of respective ones of the lens positions,
- wherein the normalised power ratio metric is generated for each corresponding lens position by:
- generating first image data representing a pair of local images of respective different focus for the lens position;
- filtering said first image data by removing a portion of high spatial frequency content from the local images in order to provide first filtered data;
- processing said first filtered data in order to remove a portion of low spatial frequency content from the local images in order to provide second filtered data;
- determining power measures for respective ones of the local images using said second filtered data by obtaining a square of each of the values of said second filtered data for the respective processed local images;
- and
- determining a ratio of said power measures for the pair of local images.
An image capture device comprising:
- a lens;
- and
- a processor:
- generating metric data for a plurality of segments of an image, the metric data representing a plurality of focus measures for each of the segments generated at a plurality of lens positions;
- processing the metric data in order to cluster the segments into at least two groups on the basis of the respective focus measures of the segments, wherein each of at least one of the two groups includes multiple ones of the segments;
- and
- determining a lens position for each of the groups corresponding to a position of true focus for the corresponding group.
An image capture device comprising:
- a lens;
- and
- a processor:
- generating metric data for a plurality of regions of an image, the data representing a plurality of focus measures for each region generated at a plurality of lens positions;
- processing the metric data in order to cluster regions into at least one group on the basis of their respective focus measures;
- and
- determining a lens position for the group corresponding to a position of true focus for the group;
- wherein the focus measures are determined using a normalised power ratio metric generated for each of respective ones of the lens positions,
- wherein the normalised power ratio metric is generated for each corresponding lens position by:
- generating first image data representing a pair of local images of respective different focus for the lens position;
- filtering said first image data by removing a portion of high spatial frequency content from the local images in order to provide first filtered data;
- processing said first filtered data in order to remove a portion of low spatial frequency content from the local images in order to provide second filtered data;
- determining power measures for respective ones of the local images using said second filtered data by obtaining a square of each of the values of said second filtered data for the respective processed local images;
- and
- determining a ratio of said power measures for the pair of local images.
Dependencies are inferred from the claim text.