Image processing method, image processing apparatus, image processing program, and image file format
View Patent ↗In a learning process, first, images having different resolutions are obtained from a target region of the subject (S 101 ). Further, the subject characteristic of the target region is obtained (S 102 ). Then, the resolution conversion rules are learned from the images having different resolutions, and those are recorded to a storage device along with the subject characteristics (S 103 ). When converting the resolutions, the resolution conversion rules learned for the corresponding subject characteristics are applied to each region of the original image so as to convert the resolutions of the original image.
1. An image processing method comprising the steps of:
(a) obtaining a plurality of images of a subject each having a different resolution, and learning a resolution conversion rule from the plurality of images for each of subject characteristics, each of the plurality of images being a visible light image;
(b) obtaining an original image and dividing the original image into regions based on the subject characteristics; and
(c) converting a resolution of the original image by applying to each of the regions obtained in step (b) the resolution conversion rule learned in step (a) regarding the subject characteristic corresponding to the region.
2. The method of claim 1 , wherein the subject characteristics include an optical characteristic independent of a shooting distance.
3. The method of claim 2 , wherein spectral reflectance or infrared spectral reflectance of each image is used as the subject characteristic.
4. The method of claim 1 , wherein the subject characteristics include a material of a surface of the subject.
5. The method of claim 1 , wherein the subject characteristics include a reflection state of a surface of the subject.
6. The method of claim 1 , wherein the resolution conversion rule describes a corresponding relation of texture feature vectors between the images having different resolutions.
7. The method of claim 1 , wherein the subject characteristics of an image include a visible light image, and wherein learning a resolution conversion rule includes obtaining a feature vector indicative of a texture of the subject for each of the plurality of images from the visible light images and determining a resolution conversion rule based on a relation between the feature vectors obtained from the visible light images.
8. An image processing method comprising the steps of:
determining a target region of a subject;
obtaining a first image and a second image with different resolutions regarding the target region, each of the first and second images being a visible light image;
learning a resolution conversion rule for relating the first and second images to each other;
obtaining a subject characteristic for the target region; and
relating the resolution conversion rule and the subject characteristic, and registering the related resolution conversion rule and subject characteristic to a storage device.
9. An image processing method comprising the steps of:
learning a resolution conversion rule for each of subject characteristics from a plurality of images of a subject, each of the plurality of images being a visible light image and having a different resolution;
obtaining an original image and region dividing information that indicates regions on the original image, the regions being divided based on subject characteristics; and
converting a resolution of the original image by applying to each of the regions indicated by the region dividing information the resolution conversion rule regarding the subject characteristic corresponding to the region.
10. An image processing apparatus comprising:
an obtaining portion operable to obtain, regarding a target region of a subject, a first and a second images having different resolutions, as well as a third image indicating a subject characteristic;
a rule learning portion which learns, from the first and second images, a resolution conversion rule for relating the first and second images to each other; and
a regionally divided image generating portion which generates, from the third image, a regionally divided image that is divided into regions based on the subject characteristic.
11. The apparatus of claim 10 , wherein the obtaining portion obtains, as the third image, a spectral image constituted with spectral information of the subject.
12. The apparatus of claim 10 , wherein the obtaining portion comprises a zoom lens and obtains the first and second images by using an optical zooming method.
13. The apparatus of claim 10 , further comprising a learning button for giving an instruction to generate the resolution conversion rule from an exterior of the apparatus.
14. The apparatus of claim 10 , further comprising a display portion for displaying the regionally divided image.