Image processing apparatus, image processing method, and program

1. An image processing apparatus comprising a memory and at least one processor which function as:

an object detection unit configured to detect an object region based on an input image;

a setting unit configured to set parameters for controlling sizes of regions to be obtained by segmenting the image at respective positions in the image, where a first parameter set for a region outside the object region is a constant and a second parameter set for a position inside the object region is determined in accordance with that position; and

a segmentation unit configured to segment the image into a plurality of regions, each of which includes adjacent similar pixels and sizes of the regions outside the object region are controlled to be approximately same by the first parameter and sizes of the regions near a boundary of the object region are controlled to be smaller than other regions by the second parameter.

2. The image processing apparatus according to claim 1 , wherein the setting unit sets an interval of cluster centers as a parameter, and

wherein the segmentation unit segments the image by clustering of pixels.

3. The image processing apparatus according to claim 1 , wherein the setting unit sets, as a parameter, a maximum size of a region to be created by repetitive merging of regions whose smallest unit is a pixel, and

wherein the segmentation unit segments the image based on a unit of the region created by the repetitive merging of regions, according to the set parameter.

4. The image processing apparatus according to claim 1 , wherein the setting unit sets a parameter for controlling likelihood of merging of regions whose smallest unit is a pixel, and

wherein the segmentation unit segments the image based on a unit of a region to be created by repetitive merging of regions, according to the set parameter.

5. The image processing apparatus according to claim 1 , wherein the setting unit sets a number of regions for each position in the image as a parameter.

6. The image processing apparatus according to claim 5 , wherein the segmentation unit segments the image in such a way that the number of regions to be obtained at a boundary area of the detected object becomes larger than regions to be obtained for other areas.

7. The image processing apparatus according to claim 6 , wherein the setting unit sets a plurality of different parameters at the boundary area of the detected object, and for controlling likelihood of merging of regions whose smallest unit is a pixel,

wherein the segmentation unit segments the image into a plurality of regions according to the parameters set by the setting unit.

8. The image processing apparatus according to claim 6 , wherein the segmentation unit performs segmentation at the boundary area of the detected object by a plurality of different methods.

9. The image processing apparatus according to claim 1 , further comprising a classification unit configured to extract feature amounts from respective regions which have been obtained by the segmentation unit, and to classify the respective regions into a plurality of classes, according to the extracted feature amounts.

10. An image processing method comprising:

detecting an object region in an input image;

setting parameters for controlling sizes of regions to be obtained by segmenting the image at respective positions in the image, where a first parameter set for a region outside the object region is a constant and a second parameter set for a position inside the object region is determined in accordance with that position; and

segmenting the image into a plurality of regions, each of which includes adjacent similar pixels and sizes of the regions outside the object region are controlled to be approximately same by the first parameter and sizes of the regions near a boundary of the object region are controlled to be smaller than other regions by the second parameter.

11. A non-transitory computer-readable storage medium storing a program that causes a computer to execute:

detecting an object region based on an input image;

setting parameters for controlling sizes of regions to be obtained by segmenting the image at respective positions in the image, where a first parameter set for a region outside the object region is a constant and a second parameter set for a position inside the object region is determined in accordance with that position; and

segmenting the image into a plurality of regions, each of which includes adjacent similar pixels and sizes of the regions outside the object region are controlled to be approximately same by the first parameter and sizes of the regions near a boundary of the object region are controlled to be smaller than other regions by the second parameter.