WO2018126435A1 - 一种3d人体模型的建立方法和系统 - Google Patents
一种3d人体模型的建立方法和系统 Download PDFInfo
- Publication number
- WO2018126435A1 WO2018126435A1 PCT/CN2017/070412 CN2017070412W WO2018126435A1 WO 2018126435 A1 WO2018126435 A1 WO 2018126435A1 CN 2017070412 W CN2017070412 W CN 2017070412W WO 2018126435 A1 WO2018126435 A1 WO 2018126435A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- human body
- model
- body model
- initial
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
Definitions
- the present invention relates to the field of graphics technologies, and in particular, to a method and system for establishing a 3D human body model.
- 3D merchandise display technology can display products in a three-dimensional way in webpages, and consumers can watch in all directions. Commodity characteristics, intuitive understanding of product information, its effect and consumers directly face the commodity is almost the same, so the 3D model is applied to all walks of life.
- the 3D model provides users with a better way to browse, display effects, attract users, and improve user attention, interaction, and participation.
- the 3D human body model is to build a three-dimensional human body model on the computer for display or secondary development to simulate real applications.
- 3D human body model generation devices which can scan human body 3D data to a computer.
- the existing 3D human body model has disadvantages in that the scanned 3D human body data is too large, and the cost of establishing the model is high. And the scanned human body cannot add bone information, so the action cannot be bound.
- a method for establishing a 3D human body model comprising:
- the step of adjusting the initial 3D human body model according to the human body database specifically includes:
- Feature point matching and labeling are performed on the initial 3D human body model corresponding to the feature points of the markers in the human body database.
- the step of acquiring any initial 3D human body model of the input and adjusting the initial 3D human body model according to the human body database further comprises:
- a 3D human body model of any of the input marked feature points is obtained, and the marked feature points of the 3D human body model are adjusted to feature points of the human body database.
- the step of collecting and storing 3D human body data specifically includes:
- the step of acquiring the feature point mark of the 3D body data 70 feature points are marked in each human body.
- the invention discloses a system for establishing a 3D human body model, comprising:
- a data processing module configured to acquire a feature point mark of the 3D human body data, and establish a human body database according to the mark of the feature point;
- the matching establishing module is configured to acquire any initial 3D human body model input, and adjust the initial 3D human body model according to the human body database to acquire a 3D human body model.
- the matching establishment module is specifically configured to:
- Feature point matching and labeling are performed on the initial 3D human body model corresponding to the feature points of the markers in the human body database.
- the matching establishment module is further configured to:
- a 3D human body model of any of the input marked feature points is obtained, and the marked feature points of the 3D human body model are adjusted to feature points of the human body database.
- the collection module is specifically configured to:
- 70 feature points are marked in each human body.
- the method for establishing a 3D human body model of the present invention comprises: collecting and storing 3D human body data; acquiring feature point marks on the 3D human body data, and establishing a human body database according to the mark of the feature points; acquiring any initial 3D human body model input, And adjusting the initial 3D human body model according to the human body database to acquire a 3D human body model.
- the human body model turns the shape of the public human body model into a target human body model. This method is cheaper, the 3D human body model is more efficient, and the acquired 3D human body model has bone information, bindable actions, and can be edited and developed twice, which greatly facilitates the use of 3D human body models. .
- the method for establishing a 3D human body model provided by the invention can reduce the threshold for secondary development and editing of the 3D human body data obtained by the current 3D human body scan, and can automatically generate human body data of various different body types by means of parametric deformation. Therefore, the cost of the 3D human body collection and the efficiency of the collection are greatly reduced, and the technical solution can be applied to the virtual fitting, and the 3D virtual human body similar to the user's body shape can be automatically generated according to the user's body shape.
- FIG. 1 is a flow chart of a method for establishing a 3D human body model according to an embodiment of the present invention
- FIG. 2 is a flow chart of another method for establishing a 3D human body model according to an embodiment of the present invention.
- FIG. 3 is a schematic diagram of a human body posture in collecting data according to an embodiment of the present invention.
- FIG. 4 is a schematic diagram of feature point marking of collected human body data according to an embodiment of the present invention.
- FIG. 5 is a schematic diagram of a human body database marked with feature points according to an embodiment of the present invention.
- FIG. 6 is a schematic diagram of feature point matching on a human body database according to an embodiment of the present invention.
- FIG. 7 is a schematic diagram of automatic marking of a human feature point according to an embodiment of the present invention.
- FIG. 8 is a schematic diagram of a 3D human body model including feature point marks according to an embodiment of the present invention.
- FIG. 9 is a schematic diagram of a 3D human body model after parametric deformation of a 3D human male model human body model according to an embodiment of the present invention.
- FIG. 10 is a schematic diagram of a system for establishing a 3D human body model according to an embodiment of the present invention.
- Computer devices include user devices and network devices.
- the user equipment or client includes but is not limited to a computer, a smart phone, a PDA, etc.; the network device includes but is not limited to a single network service.
- a server group consisting of multiple network servers or a cloud-based cloud consisting of a large number of computers or network servers.
- the computer device can operate alone to carry out the invention, and can also access the network and implement the invention through interoperation with other computer devices in the network.
- the network in which the computer device is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a VPN network, and the like.
- first means “first,” “second,” and the like may be used herein to describe the various elements, but the elements should not be limited by these terms, and the terms are used only to distinguish one element from another.
- the term “and/or” used herein includes any and all combinations of one or more of the associated listed items. When a unit is referred to as being “connected” or “coupled” to another unit, it can be directly connected or coupled to the other unit, or an intermediate unit can be present.
- a method for establishing a 3D human body model including:
- the human body data can be collected first, then the collected human body data is marked with feature points, and then the initial 3D human body model with the feature point mark is used to perform feature point matching on the input 3D human body model, thereby the initial 3D human body.
- the model performs parameter transformation to acquire the target 3D human body model and transform the shape of the public human body model into the target human body model. This method is cheaper, the 3D human body model is more efficient, and the acquired 3D human body model has bone information, bindable actions, and can be edited and developed twice, which greatly facilitates the use of 3D human body models. .
- the method for establishing a 3D human body model provided by the invention can reduce the threshold for secondary development and editing of the 3D human body data obtained by the current 3D human body scan, and can be parameterized
- the method automatically generates human body data of various body types, thereby greatly reducing the cost of 3D human body collection and improving the efficiency of collection, and the technical solution can be applied to the virtual fitting, and can automatically generate a shape similar to the user body according to the user's body shape. 3D virtual human body.
- the step of adjusting the initial 3D human body model according to the human body database specifically includes:
- Feature point matching and labeling are performed on the initial 3D human body model corresponding to the feature points of the markers in the human body database.
- the feature points marked in the human body database can be matched to the initial 3D human body model, so that the matched feature points are marked on the initial 3D human body model, the initial 3D human body model is given parameters, and the initial 3D human body model is parameterized.
- the step of acquiring any of the initial 3D human body models of the input and adjusting the initial 3D human body model according to the human body database further comprises:
- a 3D human body model of any of the input marked feature points is obtained, and the marked feature points of the 3D human body model are adjusted to feature points of the human body database.
- the parameters of the 3D human male model can be re-adjusted, the feature points in the human body model database are marked into the 3D human male model, and the 3D human male model is parameterized, thereby retaining the bone information, the binding action and Achieve secondary development.
- the step of collecting and storing 3D human body data specifically includes:
- the 3D human body data of 200 males and 200 females can be taken, which not only facilitates data collection, but also has high collection efficiency.
- the step of acquiring feature point markers for the 3D body data is marked with 70 feature points in each human body.
- 70 feature points are arranged on the human body for marking, which can cover key components of the human body, thereby more anthropomorphic and realizing in more accurate simulation of human body movements.
- other number of feature points may be set in this embodiment, for example, 80, 66, 95, 108, and the like.
- the method for establishing a 3D human body model includes:
- the database includes 200 male 3D human body models and 200 female 3D human body models, and then each feature is marked with a feature point.
- the mark may be automatically marked at a preset location, or may be manually
- the feature points are marked for each model, and manual marking is more precise.
- each human body marks 70 feature points and of course, other number of feature points can be marked. In view of the efficiency problem, 70 feature points are preferred.
- the system learns the feature point marks in the above 3D human body database, and constructs an automatic feature point marking system, which is a system capable of automatically marking any three-dimensional human body model.
- an automatic feature point marking system which is a system capable of automatically marking any three-dimensional human body model.
- an arbitrary scanned 3D mannequin is input, and then the feature point marking system is used to automatically mark the input 3D human body model.
- the 3D human body can be deformed by parameters, or the existing 3D human male model (the male mold has also been subjected to feature point marking) can be transformed into the 3D human body.
- the body shape and can retain the bone information of the 3D human male model, so that the action binding can be performed.
- the method for establishing a 3D human body model in this embodiment includes:
- Step S1 collecting human body data.
- a human body data of 200 male human bodies and 200 female human bodies can be collected by using a 3D human body scanner.
- the posture and angle of the acquisition can be seen in Figure 3.
- the human body poses T-pose the legs are naturally separated, the shoulders are flush, the arms are raised horizontally
- the scanning process is fixed, and the tights are worn. Do not wear shoes.
- Step S2 Acquire a feature point mark on the collected human body data.
- the collected 400 (200 male and female) 3D human body data can be marked with artificial feature points, and each human body marks 70 feature points.
- the number of marked feature points may also be other, such as 95, 101, and the like.
- Step S3 Construct a human body database including feature points.
- a human body database including 400 human body models (200 males and females each) and marked with feature points can be constructed.
- Step S4 Perform intelligent learning and feature point matching on the human body database.
- the human body database constructed in step S3 can be intelligently learned and feature point matching using a robot or a robot system.
- Step S5 Acquire a human body feature point automatic marking system.
- the human body feature point automatic marking system can automatically input any 3D human body model. The points are matched and marked, so that any 3D human body model input can change its parameters or re-parameterize it, thereby facilitating the operation of changing 3D mannequin actions, secondary editing and development.
- Step S6 Input an arbitrary initial 3D human body model.
- Step S7 Automatic feature point marking is performed on the input initial 3D human body model using the human body feature point automatic marking system.
- a 3D human body model including a feature point mark that is, a parameterized 3D human body model, is output by performing automatic feature point marking on the initial 3D human body model in step 6.
- Step S8 input a 3D human male model that has marked the feature points
- Step S9 According to the initial 3D human body model after the automatic feature point marking, the 3D human male model is marked with feature points to obtain a 3D human body model.
- the 3D human body model input in step S8 is parametrically deformed, thereby acquiring a 3D human body model.
- the latest 3D human body model is output, and the body shape is the same as the 3D human body shape input in step S6, but the bone information is included, and the action can be bound and the secondary development can be realized.
- a system for establishing a 3D human body model including:
- the collecting module 201 is configured to collect and store 3D human body data
- the data processing module 202 is configured to acquire a feature point mark of the 3D human body data, and establish a human body database according to the mark of the feature point;
- the matching establishing module 203 is configured to acquire any initial 3D human body model input, and adjust the initial 3D human body model according to the human body database to acquire a 3D human body model.
- the human body data can be collected first, then the collected human body data is marked with feature points, and then the initial 3D human body model with the feature point mark is used to perform feature point matching on the input 3D human body model, thereby the initial 3D human body.
- the model performs parameter transformation to acquire the target 3D human body model and transform the shape of the public human body model into the target human body model. This method is cheaper, the 3D human body model is more efficient, and the acquired 3D human body model has bone information, bindable actions, and can be edited and developed twice, which greatly facilitates the use of 3D human body models. .
- the method for establishing a 3D human body model provided by the invention can reduce the threshold for secondary development and editing of the 3D human body data obtained by the current 3D human body scan, and can automatically generate human body data of various body types by means of parametric deformation. , which greatly reduces the acquisition of 3D human body
- the present and the collection efficiency is improved, and the technical solution can be applied to the virtual fitting, and the 3D virtual human body similar to the user's body shape can be automatically generated according to the user's body type.
- the acquisition module 201 can perform scanning acquisition using a 3D scanner.
- the matching establishment module is specifically configured to:
- Feature point matching and labeling are performed on the initial 3D human body model corresponding to the feature points of the markers in the human body database.
- the feature points marked in the human body database can be matched to the initial 3D human body model, so that the matched feature points are marked on the initial 3D human body model, the initial 3D human body model is given parameters, and the initial 3D human body model is parameterized.
- the match establishing module is further configured to:
- a 3D human body model of any of the input marked feature points is obtained, and the marked feature points of the 3D human body model are adjusted to feature points of the human body database.
- the parameters of the 3D human male model can be re-adjusted, the feature points in the human body model database are marked into the 3D human male model, and the 3D human male model is parameterized, thereby retaining the bone information, the binding action and Achieve secondary development.
- the acquisition module is specifically configured to:
- the 3D human body data of 200 males and 200 females can be taken, which not only facilitates data collection, but also has high collection efficiency.
- 70 feature points are marked in each human body.
- 70 feature points are arranged on the human body for marking, which can cover key components of the human body, thereby more anthropomorphic and realizing in more accurate simulation of human body movements.
- other number of feature points may be set in this embodiment, for example, 80, 66, 95, 108, and the like.
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Computer Graphics (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Processing Or Creating Images (AREA)
Abstract
一种3D人体模型的建立方法，包括：采集并储存3D人体数据（S101）；获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库（S102）；获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型（S103）。该方式成本更低，建立3D人体模型的效率更高，可降低目前的3D人体扫描所得到的3D人体数据提供二次开发和编辑的门槛，并且可以用参数化变形的方式自动生成各种不同体型的人体数据，从而极大降低3D人体采集的成本和提高采集的效率，并且该技术方案可以应用到虚拟试衣，能够根据用户的体型自动生成与用户体型类似的3D虚拟人体。
Description
本发明涉及图形技术领域，尤其涉及一种3D人体模型的建立方法和系统。
随着计算机软硬件技术的发展，图像图形处理能力的加强，市场上建立了越来越多的3D模型，3D商品展示技术可以在网页中将商品以立体方式交互展示，消费者可以全方位观看商品特征，直观地了解商品信息，其效果和消费者直接面对商品相差无几，因此3D模型应用到了各行各业中。3D模型给用户提供更好的浏览方式、展示效果、吸引用户，提高用户的关注、互动、参与。
其中，3D人体模型是在计算机上建立立体的人体模型，以供展示或进行二次开发，模拟真实的应用。目前有多种3D人体模型的生成设备，可以将真人的人体3D数据扫描到计算机，然而现有的3D人体模型的建立存在缺点，所扫描的3D人体数据量过大，建立模型的成本较高，并且扫描的人体无法添加骨骼信息，从而无法绑定动作。
因此，如何提供一种成本较低、可自由变形的3D人体模型的建立方法和系统，成为本领域亟需解决的问题。
发明内容
本发明的目的是提供一种成本较低、可自由变形的3D人体模型的建立方法和系统。
本发明的目的是通过以下技术方案来实现的：
一种3D人体模型的建立方法，包括：
采集并储存3D人体数据；
获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库；
获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型。
优选的，所述根据所述人体数据库对所述初始3D人体模型进行调整的步骤具体包括：
根据所述人体数据库中的标记的特征点对应的在初始3D人体模型上进行特征点匹配和标记。
优选的，所述获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整的步骤之后进一步包括：
获取输入的任一已标记特征点的3D人体公模，将该3D人体公模的已标记特征点调整为人体数据库的特征点。
优选的，所述采集并储存3D人体数据的步骤具体包括：
采集至少两个男性和两个女性的3D人体数据并储存。
优选的，所述获取对该3D人体数据的特征点标记的步骤中，每个人体中标记有70个特征点。
本发明公开一种3D人体模型的建立系统，包括：
采集模块，用于采集并储存3D人体数据；
数据处理模块，用于获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库；
匹配建立模块，用于获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型。
优选的，所述匹配建立模块具体用于：
根据所述人体数据库中的标记的特征点对应的在初始3D人体模型上进行特征点匹配和标记。
优选的，所述匹配建立模块进一步用于：
获取输入的任一已标记特征点的3D人体公模，将该3D人体公模的已标记特征点调整为人体数据库的特征点。
优选的，所述采集模块具体用于：
采集至少两个男性和两个女性的3D人体数据并储存。
优选的，所述数据处理模块获取的特征点标记中，每个人体中标记有70个特征点。
本发明的3D人体模型的建立方法由于包括：采集并储存3D人体数据；获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库；获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型。采用这种方式，就可以先采集
人体数据，然后对采集的人体数据进行特征点标记，然后用带有特征点标记的人体数据对输入的初始3D人体模型进行特征点匹配，从而对初始3D人体模型进行参数变换，获取目标的3D人体模型，将公共人体模型的外形变成目标人体模型。该方式成本更低，建立3D人体模型的效率更高，并且获取的3D人体模型拥有骨骼信息，可绑定动作，并且能够进行二次编辑和开发，极大的方便了对3D人体模型的使用。
本发明提供的3D人体模型的建立方法，可降低目前的3D人体扫描所得到的3D人体数据提供二次开发和编辑的门槛，并且可以用参数化变形的方式自动生成各种不同体型的人体数据，从而极大降低3D人体采集的成本和提高采集的效率，并且该技术方案可以应用到虚拟试衣，能够根据用户的体型自动生成与用户体型类似的3D虚拟人体。
图1是本发明实施例的一种3D人体模型的建立方法的流程图；
图2是本发明实施例的另一种3D人体模型的建立方法的流程图；
图3是本发明实施例的采集数据中人体姿势的示意图；
图4是本发明实施例的对所采集人体数据的特征点标记的示意图；
图5是本发明实施例的标记好特征点的人体数据库的示意图；
图6是本发明实施例的对人体数据库进行特征点匹配的示意图；
图7是本发明实施例的人体特征点自动标记的的示意图；
图8是本发明实施例的包含特征点标记的3D人体模型的示意图；
图9是本发明实施例的将3D人体公模人体公模进行参数化变形后的3D人体模型的示意图；
图10是本发明实施例的一种3D人体模型的建立系统的示意图。
虽然流程图将各项操作描述成顺序的处理，但是其中的许多操作可以被并行地、并发地或者同时实施。各项操作的顺序可以被重新安排。当其操作完成时处理可以被终止，但是还可以具有未包括在附图中的附加步骤。处理可以对应于方法、函数、规程、子例程、子程序等等。
计算机设备包括用户设备与网络设备。其中，用户设备或客户端包括但不限于电脑、智能手机、PDA等；网络设备包括但不限于单个网络服务
器、多个网络服务器组成的服务器组或基于云计算的由大量计算机或网络服务器构成的云。计算机设备可单独运行来实现本发明，也可接入网络并通过与网络中的其他计算机设备的交互操作来实现本发明。计算机设备所处的网络包括但不限于互联网、广域网、城域网、局域网、VPN网络等。
在这里可能使用了术语“第一”、“第二”等等来描述各个单元，但是这些单元不应当受这些术语限制，使用这些术语仅仅是为了将一个单元与另一个单元进行区分。这里所使用的术语“和/或”包括其中一个或更多所列出的相关联项目的任意和所有组合。当一个单元被称为“连接”或“耦合”到另一单元时，其可以直接连接或耦合到所述另一单元，或者可以存在中间单元。
这里所使用的术语仅仅是为了描述具体实施例而不意图限制示例性实施例。除非上下文明确地另有所指，否则这里所使用的单数形式“一个”、“一项”还意图包括复数。还应当理解的是，这里所使用的术语“包括”和/或“包含”规定所陈述的特征、整数、步骤、操作、单元和/或组件的存在，而不排除存在或添加一个或更多其他特征、整数、步骤、操作、单元、组件和/或其组合。
下面结合附图和较佳的实施例对本发明作进一步说明。
如图1所示，本实施例中公开一种3D人体模型的建立方法，包括：
S101、采集并储存3D人体数据；
S102、获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库；
S103、获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型。.
采用这种方式，就可以先采集人体数据，然后对采集的人体数据进行特征点标记，然后用带有特征点标记的人体数据对输入的初始3D人体模型进行特征点匹配，从而对初始3D人体模型进行参数变换，获取目标的3D人体模型，将公共人体模型的外形变成目标人体模型。该方式成本更低，建立3D人体模型的效率更高，并且获取的3D人体模型拥有骨骼信息，可绑定动作，并且能够进行二次编辑和开发，极大的方便了对3D人体模型的使用。
本发明提供的3D人体模型的建立方法，可以降低目前的3D人体扫描所得到的3D人体数据提供二次开发和编辑的门槛，并且可以用参数化变形
的方式自动生成各种不同体型的人体数据，从而极大降低3D人体采集的成本和提高采集的效率，并且该技术方案可以应用到虚拟试衣，能够根据用户的体型自动生成与用户体型类似的3D虚拟人体。
根据其中一个示例，所述根据所述人体数据库对所述初始3D人体模型进行调整的步骤具体包括：
根据所述人体数据库中的标记的特征点对应的在初始3D人体模型上进行特征点匹配和标记。
这样就可以将人体数据库中标记的特征点匹配到初始3D人体模型上，从而将匹配的特征点标记到初始3D人体模型上，给初始3D人体模型赋予参数，将初始3D人体模型参数化。
根据其中另一个示例，所述获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整的步骤之后进一步包括：
获取输入的任一已标记特征点的3D人体公模，将该3D人体公模的已标记特征点调整为人体数据库的特征点。
采用这种方式就可以将3D人体公模的参数重新调整，将人体模型数据库中的特征点标记到3D人体公模中，将3D人体公模参数化，从而保留骨骼信息，可绑定动作以及实现二次开发。
根据其中另一个示例，所述采集并储存3D人体数据的步骤具体包括：
采集至少两个男性和两个女性的3D人体数据并储存。
这样采集的样本更多，可获取更多的特征点，让人体数据库中的样本更加丰富，这样对获取的初始3D人体模型进行特征点匹配和标记时更加容易找到相接近的样本的3D人体数据，从而进行匹配和标记。本实施例中，可以取200个男性和200个女性的3D人体数据，不仅方便数据采集，而且采集的效率较高。
根据其中另一个示例，所述获取对该3D人体数据的特征点标记的步骤中，每个人体中标记有70个特征点。
本实施例中在人体上设置70个特征点进行标记，可覆盖人体的关键部件，从而在更加精确的模拟人体动作等，更加拟人化和真实化。当然本实施例中也可以设置其他数量的特征点，例如80个，66个，95个，108个等等。
本实施例中，在更加详细和具体的描述中，3D人体模型的建立方法包括：
首先构建3D人体数据库。本实施例中该数据库包含200个男性3D人体模型以及200个女性3D人体模型，然后对每个模型进行特征点的标记，本实施例中标记可以是在预设部位自动标记，也可以是手动对每个模型进行特征点的标记，手动标记更加精确。本实施例中每个人体标记70个特征点，当然也可以标记其他数量的特征点，考虑到效率问题，70个特征点较佳。
之后，利用机器人学习的方法，让系统学习上述3D人体数据库中的特征点标记，构建一个自动特征点标记系统，该系统是一个能够自动标记任意一个三维人体模型。本实施例中，对于输入的任一三维人体模型，
在处理过程，输入任意扫描的3D人体模型，然后利用特征点标记系统，自动给输入的3D人体模型进行特诊点标记。
最后，根据已经自动标记好特征点的3D人体，可以通过参数对该3D人体进行体型变形，或者能够将已有的3D人体公模(该公模也已经进行特征点标记)变形成该3D人体的体型，并且能够保留3D人体公模的骨骼信息，从而可以进行动作绑定。
具体的，结合图2所示，本实施例中3D人体模型的建立方法包括：
步骤S1：采集人体数据。本实施例中，可利用3D人体扫描仪，可一共采集200个男性人体以及200个女性人体的人体数据。采集的姿势和角度可参见图3所示，扫描时，人体摆出T-pose(两腿自然分开，与肩平齐，两胳膊水平抬起的姿势)，扫描过程固定姿势，穿紧身衣，不穿鞋。
步骤S2：获取对所采集人体数据的特征点标记。参见图4所示，本实施例中，可对所采集的400个(其中男女各200个)3D人体数据进行人工特征点标记，每个人体标记70个特征点。当然标记特征点的数量也可以是其他的，例如95、101个等。
步骤S3：构建包括特征点的人体数据库。本实施例中，参见图5所示，可构建一个包含400个人体模型(其中男女各200个)并且标记好特征点的人体数据库。
步骤S4：对人体数据库进行智能学习及特征点匹配。本实施例中，参见图6所示，可使用机器人或机器人系统对步骤S3中构建的人体数据库进行智能学习及特征点匹配。
步骤S5：获取人体特征点自动标记系统。本实施例中，参见图7所示，该人体特征点自动标记系统可以将输入的任一3D人体模型进行自动的特
征点匹配和标记，从而将输入的任一3D人体模型改变其参数或将其重新参数化，从而方便改变3D人体模型动作、进行二次编辑和开发等操作。
步骤S6：输入任意的一个初始3D人体模型。
步骤S7：使用人体特征点自动标记系统，对输入的初始3D人体模型进行自动特征点标记。本实施例中，参见图8，通过对步骤6中初始3D人体模型进行自动特征点标记，从而输出一个包含特征点标记的3D人体模型，即参数化的3D人体模型。
步骤S8：输入一个已经标记好特征点的3D人体公模；
步骤S9：根据上述自动特征点标记后初始3D人体模型，将3D人体公模进行特征点标记，获取到3D人体模型。本实施例中，参见图9，根据步骤S7中输出的参数化的3D人体模型，将步骤S8中输入的3D人体公模进行参数化变形，从而获取到3D人体模型。
因此，本实施例中输出最新的3D人体模型，体型跟步骤S6中输入的3D人体体型一致，但是包含了骨骼信息，可以绑定动作以及实现二次开发。
根据本发明其中的一个实施例，如图10所示，公开一种3D人体模型的建立系统，包括：
采集模块201，用于采集并储存3D人体数据；
数据处理模块202，用于获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库；
匹配建立模块203，用于获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型。
采用这种方式，就可以先采集人体数据，然后对采集的人体数据进行特征点标记，然后用带有特征点标记的人体数据对输入的初始3D人体模型进行特征点匹配，从而对初始3D人体模型进行参数变换，获取目标的3D人体模型，将公共人体模型的外形变成目标人体模型。该方式成本更低，建立3D人体模型的效率更高，并且获取的3D人体模型拥有骨骼信息，可绑定动作，并且能够进行二次编辑和开发，极大的方便了对3D人体模型的使用。
本发明提供的3D人体模型的建立方法，可以降低目前的3D人体扫描所得到的3D人体数据提供二次开发和编辑的门槛，并且可以用参数化变形的方式自动生成各种不同体型的人体数据，从而极大降低3D人体采集的成
本和提高采集的效率，并且该技术方案可以应用到虚拟试衣，能够根据用户的体型自动生成与用户体型类似的3D虚拟人体。本实施例中采集模块201可以采用3D扫描仪进行扫描采集。
根据其中一个示例，所述匹配建立模块具体用于：
根据所述人体数据库中的标记的特征点对应的在初始3D人体模型上进行特征点匹配和标记。
这样就可以将人体数据库中标记的特征点匹配到初始3D人体模型上，从而将匹配的特征点标记到初始3D人体模型上，给初始3D人体模型赋予参数，将初始3D人体模型参数化。
根据其中另一个示例，所述匹配建立模块进一步用于：
获取输入的任一已标记特征点的3D人体公模，将该3D人体公模的已标记特征点调整为人体数据库的特征点。
采用这种方式就可以将3D人体公模的参数重新调整，将人体模型数据库中的特征点标记到3D人体公模中，将3D人体公模参数化，从而保留骨骼信息，可绑定动作以及实现二次开发。
根据其中另一个示例，所述采集模块具体用于：
采集至少两个男性和两个女性的3D人体数据并储存。
这样采集的样本更多，可获取更多的特征点，让人体数据库中的样本更加丰富，这样对获取的初始3D人体模型进行特征点匹配和标记时更加容易找到相接近的样本的3D人体数据，从而进行匹配和标记。本实施例中，可以取200个男性和200个女性的3D人体数据，不仅方便数据采集，而且采集的效率较高。
根据其中另一个示例，所述数据处理模块获取的特征点标记中，每个人体中标记有70个特征点。
本实施例中在人体上设置70个特征点进行标记，可覆盖人体的关键部件，从而在更加精确的模拟人体动作等，更加拟人化和真实化。当然本实施例中也可以设置其他数量的特征点，例如80个，66个，95个，108个等等。
以上内容是结合具体的优选实施方式对本发明所作的进一步详细说明，不能认定本发明的具体实施只局限于这些说明。对于本发明所属技术领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干简单推演或替换，都应当视为属于本发明的保护范围。
Claims (10)
- 一种3D人体模型的建立方法，其特征在于，包括：采集并储存3D人体数据；获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库；获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型。
- 根据权利要求1所述的方法，其特征在于，所述根据所述人体数据库对所述初始3D人体模型进行调整的步骤具体包括：根据所述人体数据库中的标记的特征点对应的在初始3D人体模型上进行特征点匹配和标记。
- 根据权利要求1所述的方法，其特征在于，所述获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整的步骤之后进一步包括：获取输入的任一已标记特征点的3D人体公模，将该3D人体公模的已标记特征点调整为人体数据库的特征点。
- 根据权利要求1所述的方法，其特征在于，所述采集并储存3D人体数据的步骤具体包括：采集至少两个男性和两个女性的3D人体数据并储存。
- 根据权利要求1所述的方法，其特征在于，所述获取对该3D人体数据的特征点标记的步骤中，每个人体中标记有70个特征点。
- 一种3D人体模型的建立系统，其特征在于，包括：采集模块，用于采集并储存3D人体数据；数据处理模块，用于获取对该3D人体数据的特征点标记，并根据特征点的标记建立人体数据库；匹配建立模块，用于获取输入的任一初始3D人体模型，并根据所述人体数据库对所述初始3D人体模型进行调整，获取3D人体模型。
- 根据权利要求6所述的系统，其特征在于，所述匹配建立模块具体用于：根据所述人体数据库中的标记的特征点对应的在初始3D人体模型上进行特征点匹配和标记。
- 根据权利要求6所述的系统，其特征在于，所述匹配建立模块进一 步用于：获取输入的任一已标记特征点的3D人体公模，将该3D人体公模的已标记特征点调整为人体数据库的特征点。
- 根据权利要求6所述的系统，其特征在于，所述采集模块具体用于：采集至少两个男性和两个女性的3D人体数据并储存。
- 根据权利要求6所述的系统，其特征在于，所述数据处理模块获取的特征点标记中，每个人体中标记有70个特征点。
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201780061461.1A CN109891464B (zh) | 2017-01-06 | 2017-01-06 | 一种3d人体模型的建立方法和系统 |
| PCT/CN2017/070412 WO2018126435A1 (zh) | 2017-01-06 | 2017-01-06 | 一种3d人体模型的建立方法和系统 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/CN2017/070412 WO2018126435A1 (zh) | 2017-01-06 | 2017-01-06 | 一种3d人体模型的建立方法和系统 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2018126435A1 true WO2018126435A1 (zh) | 2018-07-12 |
Family
ID=62788971
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2017/070412 Ceased WO2018126435A1 (zh) | 2017-01-06 | 2017-01-06 | 一种3d人体模型的建立方法和系统 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN109891464B (zh) |
| WO (1) | WO2018126435A1 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114851565A (zh) * | 2022-04-12 | 2022-08-05 | 深圳市广德教育科技股份有限公司 | 一种使用3d打印技术制造试衣模特的方法 |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110838179B (zh) * | 2019-09-27 | 2024-01-19 | 深圳市三维人工智能科技有限公司 | 基于体测数据的人体建模方法、装置及电子设备 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102178530A (zh) * | 2011-01-24 | 2011-09-14 | 天津大学 | 基于三维点云数据的人体尺寸自动测量方法 |
| US20140225888A1 (en) * | 2011-03-14 | 2014-08-14 | Belcurves Sarl | Virtual try on simulation service |
| CN104679831A (zh) * | 2015-02-04 | 2015-06-03 | 腾讯科技（深圳）有限公司 | 一种匹配人体模型的方法及装置 |
| US20150287242A1 (en) * | 2014-04-03 | 2015-10-08 | Electronics And Telecommunications Research Institute | Apparatus and method of reconstructing 3d clothing model |
| CN105551079A (zh) * | 2015-12-18 | 2016-05-04 | 朝克 | 一种个性化尺寸自调的3d人体模特的虚拟建模方法 |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106296809B (zh) * | 2016-08-11 | 2019-02-26 | 湖南化身科技有限公司 | 一种组件式三维人像建模方法 |
-
2017
- 2017-01-06 WO PCT/CN2017/070412 patent/WO2018126435A1/zh not_active Ceased
- 2017-01-06 CN CN201780061461.1A patent/CN109891464B/zh active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102178530A (zh) * | 2011-01-24 | 2011-09-14 | 天津大学 | 基于三维点云数据的人体尺寸自动测量方法 |
| US20140225888A1 (en) * | 2011-03-14 | 2014-08-14 | Belcurves Sarl | Virtual try on simulation service |
| US20150287242A1 (en) * | 2014-04-03 | 2015-10-08 | Electronics And Telecommunications Research Institute | Apparatus and method of reconstructing 3d clothing model |
| CN104679831A (zh) * | 2015-02-04 | 2015-06-03 | 腾讯科技（深圳）有限公司 | 一种匹配人体模型的方法及装置 |
| CN105551079A (zh) * | 2015-12-18 | 2016-05-04 | 朝克 | 一种个性化尺寸自调的3d人体模特的虚拟建模方法 |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114851565A (zh) * | 2022-04-12 | 2022-08-05 | 深圳市广德教育科技股份有限公司 | 一种使用3d打印技术制造试衣模特的方法 |
| CN114851565B (zh) * | 2022-04-12 | 2024-03-29 | 深圳市广德教育科技股份有限公司 | 一种使用3d打印技术制造试衣模特的方法 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109891464B (zh) | 2023-03-07 |
| CN109891464A (zh) | 2019-06-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104978762B (zh) | 服装三维模型生成方法及系统 | |
| Zhu et al. | An efficient human model customization method based on orthogonal-view monocular photos | |
| Chen et al. | Deformable model for estimating clothed and naked human shapes from a single image | |
| JP7596015B2 (ja) | 画像処理方法、装置、及び電子機器 | |
| CN103268629B (zh) | 三维人体形态和姿态的无标记点实时复原方法 | |
| CN109448099A (zh) | 画面的渲染方法、装置、存储介质及电子装置 | |
| US20210375045A1 (en) | System and method for reconstructing a 3d human body under clothing | |
| CN112785723B (zh) | 一种基于二维服装图像与三维人体模型的自动化服装建模方法 | |
| Song et al. | 3D Body Shapes Estimation from Dressed‐Human Silhouettes | |
| CN115578393B (zh) | 关键点检测方法、训练方法、装置、设备、介质及产品 | |
| WO2024109522A1 (zh) | 图像处理方法、装置及设备 | |
| WO2018126435A1 (zh) | 一种3d人体模型的建立方法和系统 | |
| CN103942405B (zh) | 三维医学数据仿真模拟的远程共享与交互研讨方法 | |
| CN114529649A (zh) | 图像处理方法和装置 | |
| CN115936796A (zh) | 一种虚拟换妆方法、系统、设备和存储介质 | |
| CN110069829B (zh) | 一种传统体育类非物质文化遗产的保护方法 | |
| CN114494007A (zh) | 一种文本指导的自然图像超分辨率重构方法 | |
| CN118071955A (zh) | 一种基于Three.JS实现三维地图埋点方法 | |
| CN116071485B (zh) | 一种人体三维重建方法及设备 | |
| Hu et al. | Self-perceptual generative adversarial network for synthetic aperture sonar image generation | |
| CN206282360U (zh) | 博物馆临时展览规划设计的系统 | |
| CN116403243A (zh) | 一种多视角多人3d姿态估计方法、装置及设备 | |
| ZHENGDONG et al. | Real-time markerless facial motion capture of personalized 3D real human research | |
| CN115239865A (zh) | 模型构建方法、模型构建装置、介质及电子设备 | |
| Wang et al. | Personalized Hand Modeling from Multiple Postures with Multi‐View Color Images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 17889878 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 05.11.2019) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 17889878 Country of ref document: EP Kind code of ref document: A1 |