Method and/or system for reconstructing from images a personalized 3D human body model and thereof

Patent metadata

Field	Value
Publication number	US-10832472-B2
Application number	US-201816166287-A
Country	US
Kind code	B2
Filing date	Oct 22, 2018
Priority date	Oct 22, 2018
Publication date	Nov 10, 2020
Grant date	Nov 10, 2020

Abstract

A human body modelling method and apparatus and an electronic device. The method comprises: acquiring at least one human body image; extracting an original profile of a human body from the human body image; calculating a body surface profile of the human body according to the original profile; and constructing a three dimensional human body model according to the body surface profile. The establishment of a three dimensional human body model originates from two dimensional human body images, without using a human body scanner to scan a human body, which reduces the cost of establishing the three dimensional human body model. Moreover, there is no requirement for the position and dressing of a human body object, and the process of establishing the three dimensional human body model is more convenient.

First claim

The invention claimed is:

1. A method of modelling a body of a human subject using at least one or more electronic devices comprising the steps of:
acquiring at least two two-dimensional images of the body of the human subject by an image acquisition means of the at least one or more electronic devices, wherein said acquired two-dimensional images are acquired from different relative positions of the image acquisition means to the human subject;
automatically extracting using a processor of the at least one or more electronic devices two dimensional original profiles of the body of the human subject from each two dimensional acquired image;
generating using a processor of the at least one or more electronic devices a two-dimensional body surface profile of the body of the human subject for each two dimensional original profile extracted from the acquired images by:
(i) automatically determining a plurality of body boundary mark features on the two dimensional original profile by comparing one or more shape features of a point or region on the original two dimensional profile against corresponding shape features which comprise body boundary mark features on a reference two dimensional profile selected according to a human body parameter from a database of reference body profiles, wherein said human body parameter comprising the area enclosed by the original profile, the height of the human body or the weight of the human body;
(ii) selecting one or more two dimensional body profiles from a library, by calculating the difference between the shape features of the determined plurality of body boundary mark features of the original two dimensional profile and the corresponding shape features of body profiles in said library;
(iii) interpolating the selected body profiles to generate a two-dimensional body surface profile of the body of the human subject body of the human subject; and
constructing using a processor of the at least one or more electronic devices a three-dimensional human body model from each of the interpolated two-dimensional body surface profiles of the human subject.

2. The method of claim 1 , wherein after constructing a three dimensional human body model according to the body surface profiles, the method further comprises the steps of:
determining geometric positions of key points or key features on the three dimensional human body model using a processor of the at least one or more electronic devices; and
calculating using a processor of the at least one or more electronic devices the measurement data of the three dimensional human body model according to the geometric positions of the key points or the key features.

3. The method of claim 1 wherein after calculating measurement data of the three dimensional human body model, the method further comprises the steps of:
converting using a processor of the at least one or more electronic devices the measurement data according to a human body size standard.

4. The method of claim 1 wherein the step of extracting the original profile from each acquired image comprises:
when a user operation is detected on the acquired image displayed on the electronic device, taking a point or a line or a plane, at which the position for the operation is located, as a part of the original profile.

5. The method of claim 1 , wherein the step of extracting the original profile from each acquired image comprises:
extracting, in combination with an image segmentation algorithm, the original profile from the acquired image when a foreground and a background selected by a user are detected on the acquired image.

6. The method of claim 1 , wherein the steps of extracting the original profile from each acquired image comprises:
selecting a first human-like template and a second human-like template, with the size of the first human-like template being less than that of the second human-like template;
calculating using a processor, a region between the second human-like template and the first human-like template to serve as a possible region where the original profile is located; and
extracting the original profile from the possible region by using image segmentation.

7. The method of claim 6 , wherein before the step of calculating a region where the second human-like template is larger than the first human-like template, the method further comprises the steps of:
identifying using a processor a reference mark feature in the acquired image; and
adjusting the first human-like template and the second human-like template based on the position of the reference mark feature.

8. The method of claim 7 , wherein the step of adjusting the human-like templates comprises the steps of:
calculating a pixel distance between the face and feet in the acquired image; and
scaling the first human-like template and the second human-like template based on the pixel distance.

9. The method of claim 7 , wherein the step of adjusting the human-like templates further comprises the steps of:
calculating an angle between an arm and the body in the acquired image, and correspondingly adjusting the positions of arms in the first human-like template and the second human-like template; and
calculating an angle between legs in the acquired image, and correspondingly adjusting the positions of legs in the first human-like template and the second human-like template.

10. The method of claim 1 wherein the shape feature comprises at least one of:
a relative coordinate distance between the point or the region on the profile and a reference point;
the curvature of the point or the region on the profile;
or a distance between the point or the region on the profile and its corresponding point or region.

11. The method of claim 1 , wherein the step of
comparing one or more shape features of a point or region on the original two dimensional profile against corresponding shape features which comprise body boundary mark features on the selected reference two dimensional profile comprises:
determining that the shape features of the point or the region on the original profile serve as the body boundary mark feature if a difference value between the first shape feature and the second shape feature on the selected reference two dimensional profile is less than a pre-set threshold value.

12. The method of claim 1 , wherein said step of selecting one or more body profiles from the body profile library according to the shape feature of the body boundary mark feature comprises the steps of:
calculating a shape feature of a point or a region, on a body profile in the body profile library, corresponding to the body boundary mark feature to serve as a third shape feature;
calculating a difference value between the third shape feature and the shape feature of the body boundary mark feature; and
selecting one or more body profiles with the smallest difference value.

13. The method of claim 1 , wherein the step of constructing the three-dimensional human body model according to the body surface profile comprises the steps of:
constructing a three-dimensional human body figure frame according to the body surface profile; and
constructing the three-dimensional human body model based on the three-dimensional human body figure frame using a processor.

14. The method of claim 13 , wherein the step of constructing a three-dimensional human body figure frame according to the body surface profile comprises the steps of:
calculating using a processor, a plurality of morphological feature data of sections of a human body in different positions according to the body surface profile;
constructing the shapes of the sectio

Assignees

• Univ Hong Kong Polytechnic

Inventors

• Mok P Y Tracy
• Zhu Shuaiyin

Classifications

• G06V40/103

  Static body considered as a whole, e.g. static pedestrian or occupant recognition · CPC title

• G06V20/64

  Three-dimensional [3D] objects · CPC title

• G06V10/44

  Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components · CPC title

• G06T19/20 Primary

  Editing of three-dimensional [3D] images, e.g. changing shapes or colours, aligning objects or positioning parts · CPC title

• G06T17/20 Primary

  Finite element generation, e.g. wire-frame surface description, {tesselation} · CPC title