Back to results
US 2019/0378277 A1Published application
Method and device for acquiring figure parameters of a user
AssigneesBeijing Jingdong Shangke Information Technology Co., Ltd., Beijing Jingdong Century Trading Co., Ltd.
InventorGang ZHAO
Life of this patent
Estimating the term…
Abstract
The present disclosure discloses a method and a device for acquiring figure parameters of a user. The method includes the following steps: acquiring a photo of the user; processing the photo to generate a profile of the user; taking a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model; and taking figure parameters of the target parameterized three-dimensional human body model as figure parameters of the user.
Claims21
1:
- A method for acquiring figure parameters of a user, comprising:
- acquiring a photo of the user;
- processing the photo to generate a profile of the user;
- taking a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model;
- and
- taking figure parameters of the target parameterized three-dimensional human body model as figure parameters of the user.
3 more elements
2:
- The method of claim 1, wherein taking a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model comprises:
- acquiring a profile posture of the user and a profile height of the user according to the profile of the user;
- adjusting the posture of the parameterized three-dimensional human body model to be consistent with the profile posture of the user;
- adjusting the height of the parameterized three-dimensional human body model to be consistent with the profile height of the user;
- and
- taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as the target parameterized three-dimensional human body model.
3 more elements
3:
- The method of claim 2, wherein the profile posture of the user comprises an arm unfolding angle and a leg separation angle;
- acquiring the profile posture of the user according to the profile of the user comprises:
- acquiring a profile center line of the profile of the user;
- and
- acquiring the arm unfolding angle and the leg separation angle according to the profile center line.
2 more elements
4:
- The method of claim 2, wherein taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as a target parameterized three-dimensional human body model comprises:
- randomly generating N groups of figure parameters, and calculating the profile error between the profile of the parameterized three-dimensional human body model corresponding to each group of figure parameters and the profile of the user, wherein N is a natural number greater than 2;
- determining whether the profile error of a group of figure parameters with a minimum profile error is smaller than the preset value;
- and
- in the case where the profile error of the group of figure parameters with the minimum profile error is smaller than the preset value, taking the parameterized three-dimensional human body model corresponding to the group of figure parameters with the minimum profile error as the target parameterized three-dimensional human body model.
2 more elements
5:
- The method of claim 4, wherein taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as a target parameterized three-dimensional human body model further comprises:
- in the case where the profile error of the group of figure parameters with a minimum profile error is not smaller than the preset value, determining two groups of figure parameters with minimum profile errors;
- and
- combining the two groups of figure parameters with minimum profile errors to generate new N groups of figure parameters, and then executing a step of calculating a profile error between the parameterized three-dimensional human body model profile corresponding to each group of figure parameters and the profile of the user.
1 more element
6:
- The method of claim 5, wherein combining the two groups of figure parameters with minimum profile errors to generate new N groups of figure parameters comprises:
- generating new N groups of figure parameters according to an average value of each parameter in the two groups of figure parameters with minimum profile errors and a random offset value in a preset proportion range.
7:
- The method of claim 1, wherein processing the photo to generate a profile of the user comprises:
- separating a foreground from a background of the human body of the user;
- generating a binary image according to the foreground and the background of the human body of the user, wherein values of the foreground and the background of the human body of the user are different;
- and
- extracting edges of the binary image to generate a profile of the user.
2 more elements
8:
- The method of claim 7, wherein separating the foreground from the background of the human body of the user comprises:
- receiving foreground identification points and background identification points of the human body input by the user;
- and
- separating the foreground from the background of the human body of the user according to the foreground identification points and the background identification points of the human body.
1 more element
9:
- The method of claim 1, wherein after taking a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model, the method further comprises:
- measuring figure parameters of the target parameterized three-dimensional human body model.
10-20. (canceled)
21:
- A device for acquiring figure parameters of a user, including a processor;
- and
- memory storing instructions that, when executed by the processor, cause the device to:
- acquire a photo of the user;
- process the photo to generate a profile of the user;
- take a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model;
- and
- take figure parameters of the target parameterized three-dimensional human body model as figure parameters of the user.
5 more elements
22:
- The device of claim 21, wherein taking a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model comprises:
- acquiring a profile posture of the user and a profile height of the user according to the profile of the user;
- adjusting the posture of the parameterized three-dimensional human body model to be consistent with the profile posture of the user;
- adjusting the height of the parameterized three-dimensional human body model to be consistent with the profile height of the user;
- and
- taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as the target parameterized three-dimensional human body model.
3 more elements
23:
- The device of claim 22, wherein the profile posture of the user comprises an arm unfolding angle and a leg separation angle;
- acquiring the profile posture of the user according to the profile of the user comprises:
- acquiring a profile center line of the profile of the user;
- and
- acquiring the arm unfolding angle and the leg separation angle according to the profile center line.
2 more elements
24:
- The device of claim 22, wherein taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as a target parameterized three-dimensional human body model comprises:
- randomly generating N groups of figure parameters, and calculating the profile error between the profile of the parameterized three-dimensional human body model corresponding to each group of figure parameters and the profile of the user, wherein N is a natural number greater than 2;
- determining whether the profile error of a group of figure parameters with a minimum profile error is smaller than the preset value;
- and
- in the case where the profile error of the group of figure parameters with the minimum profile error is smaller than the preset value, taking the parameterized three-dimensional human body model corresponding to the group of figure parameters with the minimum profile error as the target parameterized three-dimensional human body model.
2 more elements
25:
- The device of claim 24, wherein taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as a target parameterized three-dimensional human body model further comprises:
- in the case where the profile error of the group of figure parameters with a minimum profile error is not smaller than the preset value, determining two groups of figure parameters with minimum profile errors;
- and
- combining the two groups of figure parameters with minimum profile errors to generate new N groups of figure parameters, and then executing a step of calculating a profile error between the parameterized three-dimensional human body model profile corresponding to each group of figure parameters and the profile of the user.
1 more element
26:
- The device of claim 25, wherein combining the two groups of figure parameters with minimum profile errors to generate new N groups of figure parameters comprises:
- generating new N groups of figure parameters according to an average value of each parameter in the two groups of figure parameters with minimum profile errors and a random offset value in a preset proportion range.
27:
- A computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions that, when executed by a processor, cause the device for acquiring figure parameters of a user to:
- acquire a photo of the user;
- process the photo to generate a profile of the user;
- take a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model;
- and
- take figure parameters of the target parameterized three-dimensional human body model as figure parameters of the user.
3 more elements
28:
- The computer-readable storage medium of claim 27, wherein taking a parameterized three-dimensional human body model with a projection profile consistent with the profile of the user as a target parameterized three-dimensional human body model comprises:
- acquiring a profile posture of the user and a profile height of the user according to the profile of the user;
- adjusting the posture of the parameterized three-dimensional human body model to be consistent with the profile posture of the user;
- adjusting the height of the parameterized three-dimensional human body model to be consistent with the profile height of the user;
- and
- taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as the target parameterized three-dimensional human body model.
3 more elements
29:
- The computer-readable storage medium of claim 28, wherein the profile posture of the user comprises an arm unfolding angle and a leg separation angle;
- acquiring the profile posture of the user according to the profile of the user comprises:
- acquiring a profile center line of the profile of the user;
- and
- acquiring the arm unfolding angle and the leg separation angle according to the profile center line.
2 more elements
30:
- The computer-readable storage medium of claim 28, wherein taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as a target parameterized three-dimensional human body model comprises:
- randomly generating N groups of figure parameters, and calculating the profile error between the profile of the parameterized three-dimensional human body model corresponding to each group of figure parameters and the profile of the user, wherein N is a natural number greater than 2;
- determining whether the profile error of a group of figure parameters with a minimum profile error is smaller than the preset value;
- and
- in the case where the profile error of the group of figure parameters with the minimum profile error is smaller than the preset value, taking the parameterized three-dimensional human body model corresponding to the group of figure parameters with the minimum profile error as the target parameterized three-dimensional human body model.
2 more elements
31:
- The computer-readable storage medium of claim 30, wherein taking a parameterized three-dimensional human body model with a profile error between the parameterized three-dimensional human body model profile and the profile of the user less than a preset value as a target parameterized three-dimensional human body model further comprises:
- in the case where the profile error of the group of figure parameters with a minimum profile error is not smaller than the preset value, determining two groups of figure parameters with minimum profile errors;
- and
- combining the two groups of figure parameters with minimum profile errors to generate new N groups of figure parameters, and then executing a step of calculating a profile error between the parameterized three-dimensional human body model profile corresponding to each group of figure parameters and the profile of the user.
1 more element