Arm action identification method and apparatus and image processing device

1. An apparatus, comprising:

a processor to couple to a memory and to,

perform a detection of a human body in an image to obtain multiple key points of the human body in the image; and

identify whether an arm action of the human body is an upward action or a downward action by using a shoulder key point, an elbow key point and a wrist key point in the multiple key points.

2. The apparatus according to claim 1 , wherein the detection of the human body in the image is by using a cascade pyramid network to obtain 17 key points of the human body in the image.

3. The apparatus according to claim 1 , wherein to identify the arm action, the processor is to:

calculate a first height difference between the wrist key point and the elbow key point, or calculate a second height difference between the elbow key point and the shoulder key point;

determine that the arm action is the upward action when the first height difference or the second height difference is greater than 0, and determine that the arm action is the downward action when the first height difference, or the second height difference, is equal to or less than 0.

4. The apparatus according to claim 3 , wherein the processor is to,

calculate the first height difference according to the following formula:

dh _ we=y _wrist− y _elbow; and

calculate the second height difference according to the following formula:

dh _ es=y _elbow− y _shoulder;

where, dh_we is the first height difference, dh_es is the second height difference, and y_wrist, y_elbow and y_shoulder are respective ordinates of the wrist key point, the elbow key point and the shoulder key point.

5. The apparatus according to claim 1 , wherein the processor is to:

identify whether the arm action is a forward action by using the shoulder key point, the elbow key point and the wrist key point in the multiple key points.

6. The apparatus according to claim 5 , wherein to identify whether the arm action is the forward action, the processor is to:

calculate a ratio of a distance between the shoulder key point and the wrist key point to a sum of a distance between the shoulder key point and the elbow key point and a distance between the elbow key point and the wrist key point, or calculate a first included angle between a connecting line between the shoulder key point and the elbow key point and a connecting line between the elbow key point and the wrist key point; and

determine that the arm action is the forward action when the ratio is less than a first threshold, or when the first included angle is less than a second threshold.

7. The apparatus according to claim 6 , wherein the processor is to,

calculate the ratio according to the following formula:

R

=

dist

⁡

(

shoulder

,

wrist

)

dist

⁡

(

shoulder

,

elbow

)

+

dist

⁡

(

elbow

,

wrist

)

;

and

calculate the first included angle according to the following formula:

angel_armbend

=

arccos

⁡

(

V

→

elbow

→

shoulder

·

V

→

elbow

→

wrist



V

→

elbow

→

shoulder



·



V

→

elbow

→

wrist



)

;

where, R is the ratio, dist (k1, k2) is the distance between key points k1 and k2, k1 and k2 being the shoulder key point and the wrist key point, respectively, or being the shoulder key point and the elbow key point, respectively, or being the elbow key point and the wrist key point, respectively, angel_armbend is the first included angle, {right arrow over (V)} elbow→shoulder is a vector point from the elbow key point to the shoulder key point, and {right arrow over (V)} elbow→wrist is a vector point from the elbow key point to the wrist key point.

8. The apparatus according to claim 1 , wherein the processor is to:

identify whether the arm action is a forward action by using the shoulder key point, the elbow key point and a hip key point at a same side of the human body as the elbow key point in the multiple key points.

9. The apparatus according to claim 8 , wherein to identify whether the arm action is the forward action, the processor is to:

calculate a second included angle between a connecting line between the shoulder key point and the elbow key point and a connecting line between the shoulder key point and the hip key point;

determine that the arm action is the forward action when the second included angle is greater than a third threshold value.

10. The apparatus according to claim 9 , wherein the processor is to,

calculate the second included angle according to the following formula:

angle_ab

=

arccos

⁡

(

V

→

shoulder

→

hip

·

V

→

shoulder

→

elbow



V

→

shoulder

→

hip



·



V

→

shoulder

→

elbow



)

;

where, angle_ab is the second included angle, {right arrow over (V)} shoulder→hip is a vector point from the shoulder key point to the hip key point, and {right arrow over (V)} shoulder→elbow is a vector point from the shoulder key point to the elbow key point.

11. An image processing device, comprising:

an arm action identification apparatus including a processor to couple to a memory and to,

perform a detection of a human body in an image to obtain multiple key points of the human body in the image; and

identify whether an arm action of the human body is an upward action or a downward action by using a shoulder key point, an elbow key point and a wrist key point in the multiple key points.

12. The image processing device according to claim 11 , wherein to identify the arm action, the processor is to:

calculate a first height difference between the wrist key point and the elbow key point, or calculate a second height difference between the elbow key point and the shoulder key point;

determine that the arm action is the upward action when the first height difference, or the second height difference, is greater than 0, and determine that the arm action is the downward action when the first height difference, or the second height difference, is equal to or less than 0.

13. The image processing device according to claim 11 , wherein the processor is to:

identify whether the arm action is a forward action by using the shoulder key point, the elbow key point and the wrist key point in the multiple key points.

14. The image processing device according to claim 13 , wherein to identify whether the arm action is the forward action, the processor is to:

calculate a ratio of a distance between the shoulder key point and the wrist key point to a sum of a distance between the shoulder key point and the elbow key point and a distance between the elbow key point and the wrist key point, or calculate a first included angle between a connecting line between the shoulder key point and the elbow key point and a connecting line between the elbow key point and the wrist key point as a first included angle; and

determine the arm action is the forward action when the ratio is less than a first threshold, or when the first included angle is less than a second threshold.

15. The image processing device according to claim 11 , wherein the processor is to:

identify whether the arm action is a forward action by using the shoulder key point, the elbow key point and a hip key point at a same side of the human body as the elbow key point in the multiple key points.

16. The image processing device according to claim 15 , wherein to determine whether the arm action is the forward action, the processor is to:

calculate a second included angle between a connecting line between the shoulder key point and the elbow key point and a connecting line between the shoulder key point and the hip key point;

determine that the arm action is the forward action when the second included angle is greater than a third threshold value.