Apparatus and method for extracting object information
An apparatus for extracting object information according to one embodiment includes: a padding image generator for generating a padding image including an original image; a partial image acquirer for acquiring a plurality of partial images of the padding image; an object classification result acquirer for acquiring an object classification result for each of the plurality of partial images using an object classification model; a confidence map generator for generating a confidence map having a size the same as that of the padding image and including a confidence value on the basis of the object classification result; and an object information acquirer for acquiring information on an object in the padding image on the basis of the confidence map.
1. An apparatus for extracting object information, comprising at least one hardware processor configured to implement:
a padding image generator configured to generate a padding image including an original image;
a partial image acquirer configured to acquire a plurality of partial images of the padding image;
an object classification result acquirer configured to acquire an object classification result for each of the plurality of partial images using an object classification model;
a confidence map generator configured to generate a confidence map having a size the same as that of the padding image and including a confidence value on the basis of the object classification result; and
an object information acquirer configured to acquire information on an object in the padding image on the basis of the confidence map,
wherein the confidence map generator further configured to determine a confidence value of an area corresponding to each of the plurality of partial images within the confidence map on the basis of the object classification result,
wherein when an object is detected from a specific partial image among the plurality of partial images according to the object classification result, the confidence map generator further configured to increase the confidence value of an area corresponding to the specific partial image within the confidence map.
2. The apparatus according to claim 1 , wherein an area of the padding image except the original image is set to an arbitrary pixel value.
3. The apparatus according to claim 2 , wherein the original image is positioned at the center of the padding image.
4. The apparatus according to claim 1 , wherein the partial image acquirer further configured to extract an area included in a window having a predetermined size as a partial image of the padding image while moving the window within the padding image.
5. The apparatus according to claim 4 , wherein the window has a size the same as that of the original image.
6. The apparatus according to claim 1 , wherein the object information comprises size and position information of each of one or more objects included in the padding image.
7. The apparatus according to claim 1 , wherein the object information acquirer further configured to generate a binary image corresponding to the confidence map on the basis of the confidence value included in the confidence map.
8. The apparatus according to claim 7 , wherein the object information acquirer further configured to acquire the object information by analyzing a connected component between pixels within the binary image.
9. A method for extracting object information, comprising:
generating a padding image including an original image;
acquiring a plurality of partial images of the padding image;
acquiring an object classification result for each of the plurality of partial images using an object classification model;
generating a confidence map having a size the same as that of the padding image and including a confidence value on the basis of the object classification result; and
acquiring information on an object in the padding image on the basis of the confidence map,
wherein the generating of the confidence map comprises determining a confidence value of an area corresponding to each of the plurality of partial images within the confidence map on the basis of the object classification result,
wherein when an object is detected from a specific partial image among the plurality of partial images according to the object classification result, the generating of the confidence map comprises increasing the confidence value of an area corresponding to the specific partial image within the confidence map.
10. The method according to claim 9 , wherein an area of the padding image except the original image is set to an arbitrary pixel value.
11. The method according to claim 10 , wherein the original image is positioned at the center of the padding image.
12. The method according to claim 9 , wherein the acquiring of the plurality of partial images comprises extracting an area included in a window having a predetermined size as a partial image of the padding image while moving the window within the padding image.
13. The method according to claim 12 , wherein the window has a size the same as that of the original image.
14. The method according to claim 9 , wherein the object information comprises size and position information of each of one or more objects included in the padding image.
15. The method according to claim 9 , wherein the acquiring of the object information comprises generating a binary image corresponding to the confidence map on the basis of the confidence value included in the confidence map.
16. The method according to claim 15 , wherein the acquiring of the object information comprises acquiring the object information by analyzing a connected component between pixels within the binary image.