Person counting image processing apparatus, method, and storage medium
To improve calculation efficiency by properly reducing the number of regression regions based on a size of an object to be detected, a processing apparatus has at least one processor or circuit configured to function as: a size obtain unit configured to obtain a size of a predetermined object in the image; a set unit configured to divide the image into a plurality of regions based on the size of the predetermined object obtained by the size obtain unit, and to set regression regions for estimating a number of the predetermined object, wherein the set unit is configured to inhibit to set the regression regions smaller than a predetermined minimum size corresponding to the predetermined object, and an estimate unit configured to estimate the number of the predetermined object by performing a regression process on the regression regions set by the set unit.
1. An image processing apparatus comprising:
at least one processor or circuit configured to function as:
an image obtain unit configured to obtain an image captured by a camera;
a size obtain unit configured to obtain a size of a person in the image;
a set unit configured to divide the image into a plurality of temporary regions based on the size of the person obtained by the size obtain unit, and to set the temporary regions as temporary regression regions;
a determination unit configured to determine not to set the temporary regression region as a regression region for estimating the number of the persons in a case where the determination unit determines that a size of the temporary regression region is smaller than a predetermined minimum size, and to determine to set the temporary regression region as the regression region for estimating a number of persons in a case where the determination unit determines that the size of the temporary regression region is equal or larger than the predetermined minimum size; and
an estimate unit configured to estimate a number of the persons in the regression region, which is set by the determination unit, by using a recognition model obtained by a support vector machine or deep learning for estimating a number of persons in an image.
2. The image processing apparatus according to claim 1 , wherein the determination unit sets the regression region having the predetermined minimum size in a case where the temporary regression region is determined to be smaller than the predetermined minimum size.
3. The image processing apparatus according to claim 1 , wherein the determination unit sets the regression region having a size larger than the predetermined minimum size in a case where the temporary regression region is smaller than the predetermined minimum size.
4. The image processing apparatus according to claim 1 , wherein the estimate unit estimates the number of the persons by using the recognition model obtained by machine learning for estimating a number of persons in an image of a predetermined size to which the regression regions set by the determination unit are resized.
5. The image processing apparatus according to claim 1 , wherein the set unit set the temporary regions based on the size of the person so that a ratio between a size of the temporary regression region and the size of the person becomes a predetermined ratio.
6. The image processing apparatus according to claim 1 , wherein the size obtain unit is configured to set analyze regions in the image and to obtain the minimum size of the person in each of the analyze regions.
7. An image processing method comprising:
obtaining an image captured by a camera;
obtaining a size of a person in the image;
dividing the image into a plurality of temporary regions based on the size of the person, and for setting the temporary regions as temporary regression regions;
determining not to set the temporary regression region as a regression region for estimating the number of persons in a case where the size of the temporary regression region is determined to be smaller than a predetermined minimum size and to set the temporary regression region as the regression region for estimating a number of the persons in a case where the size of the temporary regression region is determined to be equal or larger than the predetermined minimum size; and
estimating the number of the persons in the regression region, which is set by the determining by using a recognition model obtained by a support vector machine or deep learning for estimating a number of persons in an image.
8. A non-transitory computer-readable storage medium that stores a computer program for image processing apparatus to execute:
obtaining an image captured by a camera;
obtaining a size of a person in the image;
dividing the image into a plurality of temporary regions based on the size of the person and for setting the temporary regions as temporary regression regions;
determining not to set the temporary regression region as a regression region for estimating the number of persons in a case where the size of the temporary regression region is determined smaller than a predetermined minimum size and to set the temporary regression region as the regression region for estimating a number of persons in a case where the size of the temporary regression region is determined to be equal or larger than the predetermined minimum size; and
estimating the number of the persons in the regression region, which is set by the determining by using a recognition model obtained by a support vector machine or deep learning for estimating a number of persons in an image.