Image processing apparatus and image processing method

1. An image processing method executed by an image processing apparatus, the method comprising steps of:

acquiring image data;

acquiring region of interest data for indicating a targeted region in recognition processing for an object included in the image data;

acquiring information of candidate region based on a degree of being similar to an object included in the image data;

estimating an object region in the acquired image data based on the region of interest data and the information of a candidate region;

performing a preview display of a plurality of candidate images to be cropped on a display unit by using information of the estimate object region; and

receiving a selection by a user from among the plurality of candidate images that are displayed,

wherein, on the image data, cropping that corresponds to the candidate image, for which the selection was received, is performed, and

wherein the displayed candidate images are images that the estimated object region is changed based on an aspect ratio specified by the user.

2. The image processing method according to claim 1 , wherein the candidate region is extracted by being grouped based on a similarity of a region in the image data.

3. The image processing method according to claim 1 , wherein the region of interest data is a heatmap for indicating a targeted region for a determination for classifying an object included in the image data.

4. The image processing method according to claim 1 , wherein information on a plurality of candidate regions is acquired,

and the plurality of candidate regions are each assigned with a score based on the region of interest data.

5. The image processing method according to claim 4 , wherein the object region is estimated by interpolating the plurality of candidate regions based on the assigned score.

6. The image processing method according to claim 4 , wherein the object region is estimated by weighted averaging of locations of the plurality of candidate regions.

7. The image processing method according to claim 4 , wherein a weight acquired based on the region of interest data is used for weighted averaging of locations of the plurality of candidate regions.

8. The image processing method according to claim 1 , wherein a neural network is used to recognize the object in the recognition processing,

and the region of interest data, which is generated based on an intermediate image extracted from an intermediate layer of the neural network, is acquired.

9. The image processing method according to claim 1 , wherein information on the estimated object region is stored in association with the image data.

10. The image processing method according to claim 9 , further comprising steps of:

performing image processing for the image data; and

performing cropping for the image data using the stored information on the object region.

11. An image processing apparatus comprising:

at least one computer coupled to at least one processor that causes the image processing apparatus to function as:

(a) a first acquisition unit configured to acquire image data;

(b) a second acquisition unit configured to acquire region of interest data for indicating a targeted region in recognition processing for an object included in the image data;

(c) a third acquisition unit configured to acquire information of a candidate region based on a degree of being similar to an object included in the image data;

(d) an estimation unit configured to estimate an object region in the image data acquired by the first acquisition unit based on the region of interest data and the information of a candidate region;

(e) a performing unit configured to perform a preview display of a plurality of candidate images to be cropped on a display unit by using information of the estimate object region; and

(f) a receiving unit configured to receive a selection by a user from among the plurality of candidate images that are displayed,

wherein, on the image data, cropping that corresponds to the candidate image, for which the selection was received, is performed, and

wherein the displayed candidate images are images that the estimated object region is changed based on an aspect ratio specified by the user.