Image processing apparatus for forming synthetic image and image processing method for forming synthetic image
AssigneeSeiko Epson Corporation
InventorKan Matsuda
Life of this patent
Estimating the term…
Abstract
To provide a solution by which adjustment of the depth display of the image can be easily carried out by the user at will in a technique for forming a 3-D image from plural images, from the plural feed images, one feed image is extracted as the reference feed image, with an object recognition process being carried out to extract the object region having the prescribed characteristic features. The reference feed image IL is displayed on the display unit 108 together with the markers MK indicating the object regions, and the user selects one object region. A region that is similar in image content with the selected region is detected from each other feed image, with the images being shifted so that the regions overlap each other.
Claims8
An image processing apparatus that forms a 3-D viewable synthetic image via a lenticular lens based on plural feed images having parallax from each other, the image processing apparatus comprising:
- a processor that extracts a plurality of candidate objects each of which has a prescribed characteristic feature by image recognition treatment in a reference feed image among the feed images,
- displays in a display the candidate objects and a plurality of region indicators indicating candidate regions including the candidate objects, respectively,
- receives input of one of the candidate regions from a user, who selects the one of the candidate regions in the reference feed image and assigns the one of the candidate regions as an assigned region,
- detects a corresponding region corresponding to the assigned region assigned in the reference feed image from each other feed image other than the reference feed image, and
- cuts out rectangular shaped images from the plural feed images, respectively, side by side and forms the synthetic image with a smallest parallax between the assigned region and the corresponding region.
An image processing method that forms a 3-D viewable synthetic image via a lenticular lens based on plural feed images having parallax from each other, the image processing method comprising:
- extracting a plurality of candidate objects each of which has a prescribed characteristic feature by image recognition treatment in a reference feed image among the feed images;
- displaying in a display the candidate objects and a plurality of region indicators indicating candidate regions including the candidate objects, respectively:
- receiving input of one of the candidate regions from a user, who selects the one of the candidate regions in the reference feed image and assigning the one of the candidate regions as an assigned region;
- detecting a corresponding region corresponding to the assigned region from each other feed image other than the reference feed image;
- and
- forming a synthetic image by setting side by side rectangular shaped images that are cut out from an overlapped portion where the assigned region of the reference feed image and the corresponding region of the other feed image are overlapped with each other, in the reference feed image and the other feed image.
Dependencies are inferred from the claim text.