Method and apparatus for detecting living body, electronic device, and storage medium

1. A method for detecting a living body, comprising:

performing target object detection on a first image captured by a first image sensor in a binocular camera apparatus to obtain a first target region, and performing the target object detection on a second image captured by a second image sensor in the binocular camera apparatus to obtain a second target region;

determining whether the first target region as a whole is consistent with the second target region;

determining, in response to the first target region as a whole being inconsistent with the second target region, a key region in the first target region based on key point information of the first target region, wherein the key point information comprises position information of key points in the first target region;

determining a corresponding region of the first target region in the second image based on position information of the key region;

obtaining key point depth information of a target object according to the first target region and the corresponding region; and

determining, based on the key point depth information of the target object, whether the target object is a living body.

2. The method according to claim 1 , wherein the obtaining key point depth information of the target object according to the first target region and the corresponding region comprises:

performing key point detection on the first target region to obtain key point information of the first target region, and performing the key point detection on the corresponding region to obtain key point information of the corresponding region; and

determining key point depth information of the target object according to the key point information of the first target region and the key point information of the corresponding region.

3. The method according to claim 1 , wherein the determining, based on the key point depth information of the target object, whether the target object is the living body comprises:

determining depth dispersion based on the key point depth information of the target object; and

determining, according to the depth dispersion, whether the target object is the living body.

4. The method according to claim 1 , wherein the determining whether the first target region as a whole is consistent with the second target region comprises:

searching a database for a first search result corresponding to the first target region;

searching the database for a second search result corresponding to the second target region; and

determining, based on the first search result and the second search result, whether the first target region as a whole is consistent with the second target region.

5. The method according to claim 1 , wherein the determining whether the first target region as a whole is consistent with the second target region comprises:

determining a similarity between the first target region and the second target region; and

determining, based on the similarity, whether the first target region as a whole is consistent with the second target region.

6. The method according to claim 1 , wherein obtaining the key point depth information of the target object according to the first target region and the corresponding region comprises:

determining the corresponding region as a corrected second target region;

obtaining the key point depth information of the target object according to the first target region and the corrected second target region.

7. The method according to claim 1 , wherein the determining the key region in the first target region according to the key point information of the first target region comprises:

determining, based on the key point information of the first target region, a smallest region enclosed by at least one key point in the first target region; and

amplifying the smallest region by a preset number of times to obtain the key region.

8. The method according to claim 1 , wherein the determining the corresponding region of the first target region in the second image comprises:

mapping at least one key point in the first target region to the second image to obtain mapping position information of the at least one key point in the second image; and

determining the corresponding region of the first target region in the second image according to the mapping position information of the at least one key point in the second image.

9. The method according to claim 1 , further comprising:

determining whether fake information exists in the first image and the second image in response to determining, based on the key point depth information of the target object, that the target object is the living body; and

determining, based on whether the fake information exists in the first image and the second image, whether the target object is the living body.

10. The method according to claim 9 , wherein the determining whether the fake information exists in the first image and the second image comprises:

performing feature extraction processing on the first image and the second image separately to obtain first feature data and second feature data; and

determining, based on the first feature data and the second feature data, whether the fake information exists in the first image and the second image.

11. The method according to claim 10 , wherein the determining, based on the first feature data and the second feature data, whether the fake information exists in the first image and the second image comprises:

performing fusion processing on the first feature data and the second feature data to obtain fusion features; and

determining, based on the fusion features, whether the fake information exists in the first image and the second image.

12. The method according to claim 1 , further comprising:

determining whether the first image and the second image satisfy a frame selection condition;

wherein the obtaining key point depth information of the target object according to the first target region and the second target region comprises:

obtaining, in response to determining that the first image and the second image satisfy the frame selection condition, key point depth information of the target object according to the first target region and the second target region.

13. The method according to claim 12 , wherein the frame selection condition comprises one or any combination of the following conditions:

the target object is detected in both the first image and the second image;

the target object detected in the first image is located in a set region of the first image and the target object detected in the second image is located in a set region of the second image;

completeness of the target object detected in the first image and completeness of the target object detected in the second image satisfy a preset condition;

a proportion, in the first image, of the target object detected in the first image is greater than a proportion threshold and a proportion, in the second image, of the target object detected in the second image is greater than the proportion threshold;

clarity of the first image and clarity of the second image both are greater than a clarity threshold; and

exposure of the first image and exposure of the second image both are greater than an exposure threshold.

14. The method according to claim 12 , further comprising:

determining, in response to determining that at least one of the first image or the second image does not satisfy the frame selection condition, whether a next image pair in a video stream satisfies the frame selection condition; and

determining the video stream as a fake video stream in response to determining that no image pair satisfying the frame selection condition is found from the video stream within a preset time period or within a preset number of image pairs.

15. The method according to claim 1 , wherein the target object is a human face.

16. An apparatus for detecting a living body, comprising:

a processor; and

memory for storing instructions executable by the processor;

wherein execution of the instructions by the processor causes the processor to perform operations, the operations comprising:

performing target object detection on a first image captured by a first image sensor in a binocular camera apparatus to obtain a first target region, and performing the target object detection on a second image captured by a second image sensor in the binocular camera apparatus to obtain a second target region;

determining whether the first target region as a whole is consistent with the second target region;

determining, in response to the first target region as a whole being inconsistent with the second target region, a key region in the first target region based on key point information of the first target region, wherein the key point information comprises position information of key points in the first target region;

determining a corresponding region of the first target region in the second image based on position information of the key region;

obtaining key point depth information of a target object according to the first target region and the corresponding region; and

determining, based on the key point depth information of the target object, whether the target object is a living body.

17. A non-transitory computer-readable storage medium, having computer program instructions stored thereon, wherein execution of the computer program instructions by a processor causes the processor to:

perform target object detection on a first image captured by a first image sensor in a binocular camera apparatus to obtain a first target region, and perform the target object detection on a second image captured by a second image sensor in the binocular camera apparatus to obtain a second target region;

determine whether the first target region as a whole is consistent with the second target region;

determine, in response to the first target region as a whole being inconsistent with the second target region, a key region in the first target region based on key point information of the first target region, wherein the key point information comprises position information of key points in the first target region;

determine a corresponding region of the first target region in the second image based on position information of the key region;

obtain key point depth information of a target object according to the first target region and the corresponding region; and

determine, based on the key point depth information of the target object, whether the target object is a living body.