Feature extraction apparatus, feature extraction program, and image processing apparatus
Provided is a feature extraction device including a sub region setting unit ( 433 ) which sets a plurality of sub regions with respect to a pixel of interest, and a binary pattern generation unit ( 434 ) which generates a local binary pattern indicating pixel value comparison results relative to each sub region with respect to each pixel of interest. The sub region setting unit ( 433 ) sets ( 436 ) a region constituted by a plurality of pixels including a pixel separated from the pixel of interest as a sub region, and the binary pattern generation unit ( 434 ) calculates ( 437 ) a representative value for each sub region and generates ( 439 ) a local binary pattern indicating whether or not the difference ( 438 ) between the representative value and the value of the pixel of interest is equal to or larger than a predetermined threshold value.
1. A feature extraction apparatus comprising:
a sub-region setting section that sets all or some pixels in an image taken by a camera to be pixels of interest, respectively, and that sets a plurality of sub-regions with respect to each of the pixels of interest based on spatial frequency characteristics of the camera; and
a binary pattern generating section that generates a local binary pattern for each of the pixels of interest, the local binary pattern indicating comparison results between a pixel value of the pixel of interest and pixel values of the plurality of set sub-regions using bit values, wherein
the plurality of sub-regions include at least a sub-region constituted by a plurality of pixels including a pixel separated from a corresponding one of the pixels of interest by one or more pixels, and
at least one of a width and a length of the sub-region is equal to or less than the number of pixels equivalent to the reciprocal of the maximum value of a spatial frequency of the camera.
2. The feature extraction apparatus according to claim 1 , wherein the binary pattern generating section calculates representative values respectively for the sub-regions, each of the representative values representing a pixel value of one pixel or a group of pixel values of a plurality of pixels constituting a corresponding one of the sub-regions, and generates, as the local binary pattern, bit data indicating whether differences between a pixel value of the pixel of interest and the representative values are each equal to or greater than a predetermined threshold using bit values.
3. The feature extraction apparatus according to claim 1 , wherein the sub-region setting section sets each of the sub-regions so that at least one of a width of the sub-region, a length of the sub-region, a distance of the sub-region from the pixel of interest, and a distance of the sub-region from another sub-region is set to such a size that all pixel values of the sub-region in a length direction of the sub-region are regarded as being nearly equal to each other in view of the spatial frequency characteristics of the camera.
4. The feature extraction apparatus according to claim 1 , wherein the sub-region setting section sets the plurality of sub-regions with an equal distance from the pixel of interest.
5. The feature extraction apparatus according to claim 1 , wherein the sub-region setting section sets the plurality of sub-regions so that the sub-regions do not overlap each other.
6. The feature extraction apparatus according to claim 1 , wherein the sub-region setting section sets the plurality of sub-regions so that the sub-regions are separated from each other.
7. The feature extraction apparatus according to claim 1 , wherein the sub-region setting section sets the plurality of sub-regions so that the sub-regions are separated by equiangular intervals to the pixel of interest.
8. The feature extraction apparatus according to claim 1 , wherein the sub-region setting section sets at least a plurality of pixels adjacent to the pixel of interest, and regions each constituted by a plurality of pixels separated from the pixel of interest by two or more pixels, to be the sub-regions, respectively.
9. An image processing apparatus comprising:
the feature extraction apparatus according to claim 1 , further including a histogram generating section that generates a histogram indicating a distribution of the local binary patterns generated from the image; and
an identifying section that determines whether the image includes a predetermined object using a classifier for identifying the predetermined object, based on the histogram generated by the feature extraction apparatus.
10. A feature extraction method comprising:
setting all or some pixels in an image taken by a camera to be pixels of interest, respectively, and setting a plurality of sub-regions with respect to each of the pixels of interest based on spatial frequency characteristics of the camera; and
generating a local binary pattern for each of the pixels of interest, the local binary pattern indicating comparison results between a pixel value of the pixel of interest and pixel values of the plurality of set sub-regions using bit values, wherein
the plurality of sub-regions include at least a sub-region constituted by a plurality of pixels including a pixel separated from a corresponding one of the pixels of interest by one or more pixels, and
at least one of a width and a length of the sub-region is equal to or less than the number of pixels equivalent to the reciprocal of the maximum value of a spatial frequency of the camera.