Image identification device, image identification method and recording medium
The invention provides an image identification device uses a separating plane to classify block images into the categories. The image identification device includes a target image input unit inputting a target image, a block image generation unit generates block images, a feature quantity computing unit computes feature quantities of the block images, and a category determination unit determines whether the block images are classified into the categories or not. The feature quantity computing unit uses local feature quantities of the block images and a global feature quantity of the target image as a whole, and also in a feature quantity space using features of the block images as coordinate axes, uses coordinate positions of feature quantity vectors optional areas in the feature quantity space to count the block images and causes the global feature quantity to include the number of the block images thus counted.
1. An image identification device that learns in advance a separating plane used for classifying images into predetermined categories in a first feature quantity space using a feature of the image as a coordinate axis and uses the separating plane to classify block images obtained by dividing a target image into the categories, the image identification device comprising:
an input unit that inputs the target image;
a block image generation unit that divides the target image into a plurality of blocks to generate the block images;
a feature quantity computing unit that computes feature quantities of the block images; and
a category determination unit that determines whether the block images are classified into one of the categories or not, using the separating plane and coordinate positions corresponding to magnitudes of feature quantities of the block images in the first feature quantity space,
wherein
the feature quantity computing unit uses, as feature quantities for the block images, local feature quantities calculated according to image information of the block images and a global feature quantity calculated according to image information of the target image as a whole, and
in a second feature quantity space where a plurality of features of the block images are used as coordinate axes, using coordinate positions of feature quantity vectors made by combining the local feature quantities of the block images and using at least one optional area, the number of block images that have the feature quantity vectors belonging to the area is counted on an area-by-area basis to be included in the global feature quantity.
2. The image identification device according to claim 1 , wherein the global feature quantity includes out of the block images included in the target image as a whole, the number of the block images having magnitudes of the local feature quantities equal to or larger than a predetermined value or the number of the block images having magnitudes of the local feature quantities smaller than the predetermined value.
3. The image identification device according to claim 1 , further comprising a target area image extraction unit that extracts a target area from the target image to be used as a target area image, wherein
the block image generation unit divides the target area image into a plurality of blocks to generate the block images.
4. The image identification device according to claim 2 , further comprising a target area image extraction unit that extracts a target area from the target image to be used as a target area image, wherein
the block image generation unit divides the target area image into a plurality of blocks to generate the block images.
5. An image identification method for learning in advance a separating plane used for classifying images into predetermined categories in a first feature quantity space using a feature of the image as a coordinate axis and then using the separating plane to classify block images obtained by dividing a target image into the categories, the image identification method comprising:
an input step of inputting the target image;
a block image generation step of dividing the target image into a plurality of blocks to generate the block images;
a feature quantity computing step of computing feature quantities of the block images; and
a category determination step of determining whether the block images are classified into one of the categories or not, using the separating plane and coordinate positions corresponding to magnitudes of feature quantities of the block images in the first feature quantity space,
wherein
at the feature quantity computing step,
as feature quantities for the block images, local feature quantities calculated according to image information of the block images and a global feature quantity calculated according to image information of the target image as a whole are used, and
in a second feature quantity space where a plurality of features of the block images are used as coordinate axes, using coordinate positions of feature quantity vectors made by combining the local feature quantities of the block images and using at least one optional area, the number of block images that have the feature quantity vectors belonging to the area is counted on an area-by-area basis to be included in the global feature quantity.
6. A non-transitory computer-readable recording medium in which an image identification program is stored that causes a computer to operate so as to learn in advance a separating plane used for classifying images into predetermined categories in a first feature quantity space using a feature of the image as a coordinate axis and use the separating plane to classify block images obtained by dividing a target image into the categories, the recording medium storing therein an image identification program that causes the computer to operate:
as an input unit that inputs the target image;
a block image generation unit that divides the target image into a plurality of blocks to generate the block images;
a feature quantity computing unit that computes feature quantities of the block images; and
a category determination unit that determines whether the block images are classified into one of the categories or not, using the separating plane and coordinate positions corresponding to magnitudes of feature quantities of the block images in the first feature quantity space,
wherein
the feature quantity computing unit uses, as feature quantities for the block images, local feature quantities calculated according to image information of the block images and a global feature quantity calculated according to image information of the target image as a whole, and
in a second feature quantity space where a plurality of features of the block images are used as coordinate axes, using coordinate positions of feature quantity vectors made by combining the local feature quantities of the block images and using at least one optional area, the number of block images that have the feature quantity vectors belonging to the area is counted on an area-by-area basis to be included in the global feature quantity.