Object centering method of mutiple faces and object centering device of mutiple faces

Abstract

This invention provides a method for object centering of multiple faces, wherein a processor reads at least one instruction stored in memory to execute the method. The method includes the following steps: obtaining multiple face depths based on a depth image and multiple face images; determining whether the multiple face depths are within a depth threshold range; if the depth of a first face among the multiple face depths is within the depth threshold range, obtaining a first face image from the multiple face images based on the first face depth; and centering an input image containing the first face image based on the first face image to generate an object-centered image.

Claims

1. A method for object centering of a plurality of faces, wherein a processor reads at least one instruction stored in a memory to execute the method, comprising: obtaining a plurality of face depths based on a depth image and a plurality of face images; determining whether the plurality of face depths are within a depth threshold range; if a first face depth among the plurality of face depths is within the depth threshold range, obtaining a first face image among the plurality of face images based on the first face depth; and performing object centering on an input image containing the first face image based on the first face image to generate an object-centered image.

2. The object centering method for a plurality of faces as described in claim 1, wherein centering the object based on the first face image to generate the object-centered image comprises: using the first face image as a centering reference for the object centering; and centering the object based on the centering reference on the input image containing the first face image to generate the object-centered image.

3. The object centering method for a plurality of faces as described in claim 1 further includes: if a second face depth among the plurality of face depths is not within the depth threshold range, obtaining a second face image from the plurality of face images based on the second face depth.

4. The object centering method for multiple faces as described in claim 3 further includes: not using the second face image as a centering reference for object centering; and not performing object centering on the input image containing the second face image based on the centering reference to generate the object-centered image.

5. The object centering method for multiple faces as described in claim 1 further includes: performing a depth estimation on the input image to generate the depth image.

6. The object centering method for a plurality of faces as described in claim 5, wherein the step of performing the depth estimation on the input image to generate the depth image comprises: performing a scaling down step on the input image to generate a scaled-down image; and performing the depth estimation on the scaled-down image to generate the depth image.

7. The object centering method for multiple faces as described in claim 1 further includes: performing face detection on the input image to generate the multiple face images.

8. The object centering method for a plurality of faces as described in claim 7, wherein the step of performing face detection on the input image to generate the plurality of face images comprises: performing a reduction step on the input image to generate a reduced image; and performing face detection on the reduced image to generate the plurality of face images.

9. The object centering method for a plurality of faces as described in claim 1, wherein centering the object based on the first face image to generate the object-centered image comprises: performing a box correction on the first face image to generate a corrected box; and performing an image magnification on the corrected box to generate the object-centered image.

10. An object centering device for a plurality of faces, comprising: a memory for storing at least one instruction; and a processor for reading the at least one instruction to perform the following steps: obtaining a plurality of face depths based on a depth image and a plurality of face images; determining whether the plurality of face depths are within a depth threshold range; if a first face depth among the plurality of face depths is within the depth threshold range, obtaining a first face image among the plurality of face images based on the first face depth; and performing object centering on an input image containing the first face image based on the first face image to generate an object-centered image.

Images