技术领域Technical Field
本发明涉及一种基于二维影像生成三维模型的数据处理方法及组件。The invention relates to a data processing method and a component for generating a three-dimensional model based on a two-dimensional image.
背景技术Background technique
3D摄像机，利用的是3D镜头制造的摄像机，通常具有两个摄像镜头以上，间距与人眼间距相近，能够拍摄出类似人眼所见的针对同一场景的不同图像。A 3D camera is a camera made using a 3D lens, which usually has two or more camera lenses with a spacing close to that of human eyes. It can capture different images of the same scene similar to what the human eye sees.
第一台3D摄像机迄今3D革命全部围绕好莱坞重磅大片和重大体育赛事展开。随着3D摄像机的问世，这项技术距离家庭用户又近了一步。在这款摄像机推出以后，我们今后就可以用3D镜头捕捉人生每一个难忘瞬间，比如孩子迈出的第一步，大学毕业庆典等。The first 3D camera The 3D revolution has so far been all about Hollywood blockbusters and major sporting events. With the advent of the 3D camera, the technology is one step closer to home users. With the launch of this camera, we will be able to capture every memorable moment in life in 3D, from a child's first steps to a college graduation.
3D摄像机通常有两个以上镜头。3D摄像机本身的功能就像人脑一样，可以将两个镜头图像融合在一起，变成一个3D图像。这些图像可以在3D电视上播放，观众佩戴所谓的主动式快门眼镜即可观看，也可通过裸眼3D显示设备直接观看。3D快门式眼镜能够以每秒60次的速度令左右眼镜的镜片快速交错开关。这意味着每只眼睛看到的是同一场景的稍显不同的画面，所以大脑会由此以为其是在欣赏以3D呈现的单张照片。3D cameras usually have two or more lenses. The 3D camera itself functions like the human brain, fusing the two lens images together into one 3D image. These images can be shown on a 3D TV, viewed by viewers wearing so-called active shutter glasses, or directly through a naked-eye 3D display device. 3D shutter glasses rapidly switch the lenses of the left and right glasses on and off 60 times per second. This means that each eye sees a slightly different version of the same scene, so the brain thinks it is viewing a single picture in 3D.
现在市场上的3D摄像机过于昂贵，现有技术中可以利用多张2D影像建立3D模型，但这种方式需要的2D影像数量较多，导致所需要的计算资源很高，算力成本也昂贵。The 3D cameras on the market are too expensive. The existing technology can use multiple 2D images to build a 3D model, but this method requires a large number of 2D images, resulting in high computing resources and expensive computing costs.
发明内容Summary of the invention
本发明要解决的技术问题是为了克服现有技术中3D摄像机过于昂贵，利用多张2D影像建立3D模型所需要的2D影像数量较多，导致所需要的计算资源很高，算力成本也昂贵的缺陷，提供一种能够利用三张二维影像生成3D模型，提高人脸识别的精度，且能够获取高画质的3D模型，产品支持多达90度完全侧脸时的人脸识别，在摄像头过高、在拍摄目标偏转头部时依然能快速锁定目标的基于二维影像生成三维模型的数据处理方法及组件。The technical problem to be solved by the present invention is to overcome the defects in the prior art that 3D cameras are too expensive, and a large number of 2D images are required to build a 3D model using multiple 2D images, resulting in high computing resources and expensive computing costs. A method is provided that can generate a 3D model using three two-dimensional images, improve the accuracy of face recognition, and obtain a high-quality 3D model. The product supports face recognition at up to 90 degrees of complete side face, and can quickly lock the target when the camera is too high or the target's head is deflected. The method and components for generating three-dimensional models based on two-dimensional images.
本发明是通过下述技术方案来解决上述技术问题：The present invention solves the above technical problems through the following technical solutions:
一种基于二维影像生成三维模型的数据处理方法，其特点在于，所述数据处理方法包括：A data processing method for generating a three-dimensional model based on a two-dimensional image, wherein the data processing method comprises:
获取一头部目标的三张二维影像，分别为第一影像、第二影像以及第三影像，其中，所述第一影像、第三影像与第二影像的拍摄方向的夹角取值范围均为[75，105]；Acquire three two-dimensional images of a head target, which are a first image, a second image, and a third image, wherein the angles between the shooting directions of the first image, the third image, and the second image are all in the range of [75, 105];
通过人工智能算法利用所述三张二维影像获取所述头部目标的三维模型。The three two-dimensional images are used by an artificial intelligence algorithm to obtain a three-dimensional model of the head target.
较佳地，所述数据处理方法还包括：Preferably, the data processing method further includes:
获取一三维模型数据库，所述三维模型数据库包括原始三维影像；Acquire a three-dimensional model database, wherein the three-dimensional model database includes original three-dimensional images;
对每一原始三维影像做标准化处理获取标准三维影像；Performing standardization processing on each original three-dimensional image to obtain a standard three-dimensional image;
利用全部所述标准三维模型做人工智能学习获取训练模板；Using all the standard three-dimensional models to perform artificial intelligence learning to obtain training templates;
通过所述训练模板利用所述三张二维影像获取所述头部目标的三维模型。The three two-dimensional images are used through the training template to obtain a three-dimensional model of the head target.
较佳地，所述利用全部所述标准三维模型做人工智能学习获取训练模板，包括：Preferably, the method of using all the standard three-dimensional models for artificial intelligence learning to obtain training templates includes:
对于每一标准三维模型，获取所述标准三维模型数据特征表达，并对所述数据特征表达进行统计以得到数据特征表达的平均数据还有均方差数据；For each standard three-dimensional model, obtaining the data feature expression of the standard three-dimensional model, and performing statistics on the data feature expression to obtain average data and mean square error data of the data feature expression;
利用全部所述标准三维模型的所述平均数据和所述均方差数据做人工智能学习获取训练模板，所述训练模板通过控制参数以重建出不同人脸模型。The average data and the mean square error data of all the standard three-dimensional models are used for artificial intelligence learning to obtain training templates, and the training templates are used to reconstruct different human face models by controlling parameters.
较佳地，所述通过所述训练模板利用所述三张二维影像获取所述头部目标的三维模型，包括：Preferably, the step of obtaining the three-dimensional model of the head target by using the three two-dimensional images through the training template includes:
识别第一影像、第二影像以及第三影像的轮廓特征；identifying contour features of the first image, the second image, and the third image;
通过所述训练模板利用所述轮廓特征获取所述头部目标的三维模型。The three-dimensional model of the head target is acquired by using the contour feature through the training template.
较佳地，所述通过所述训练模板利用所述轮廓特征获取所述头部目标的三维模型，包括：Preferably, the acquiring the three-dimensional model of the head target by using the contour feature through the training template includes:
根据所述轮廓特征获取二维影像的拍摄方向；Acquire the shooting direction of the two-dimensional image according to the contour feature;
按所述拍摄方向查找与所述轮廓特征匹配的训练模板；Searching for a training template matching the contour feature according to the shooting direction;
根据所述轮廓特征调整所述训练模板的形状以获取所述头部目标的三维模型。The shape of the training template is adjusted according to the contour feature to obtain a three-dimensional model of the head target.
较佳地，所述数据处理方法包括一标准化模型，所述标准化模型为张量模型，所述对每一原始三维影像做标准化处理获取标准三维影像，包括：Preferably, the data processing method includes a standardization model, which is a tensor model. The standardization process for each original three-dimensional image to obtain a standard three-dimensional image includes:
对于每一原始三维影像，按所述原始三维模型调节所述标准化模型以获取所述标准三维影像。For each original 3D image, the standardized model is adjusted according to the original 3D model to obtain the standardized 3D image.
较佳地，所述按所述原始三维模型调节所述标准化模型以获取所述标准三维影像，包括：Preferably, adjusting the standardized model according to the original three-dimensional model to obtain the standard three-dimensional image comprises:
识别原始三维影像的五官特征点；Identify facial feature points of the original 3D image;
根据所述五官特征点调节所述标准化模型的尺寸，并将原始三维影像与标准化模型在空间中对齐；Adjusting the size of the standardized model according to the facial feature points, and aligning the original three-dimensional image with the standardized model in space;
对于标准化模型上的每一影像点，获取影像点处法线与原始三维影像的交点；For each image point on the standardized model, obtain the intersection point of the normal line at the image point and the original three-dimensional image;
按预设规则调节所述标准化模型以获取所述标准三维影像，所述预设规则为将每一影像点与对应交点的之间的长度调节为相同长度。The standardized model is adjusted according to a preset rule to obtain the standard three-dimensional image, and the preset rule is to adjust the length between each image point and the corresponding intersection point to the same length.
本发明还提供一种基于二维影像生成三维模型的数据处理组件，其特点在于，所述数据处理组件包括一获取模块以及一计算模块，The present invention also provides a data processing component for generating a three-dimensional model based on a two-dimensional image, wherein the data processing component comprises an acquisition module and a calculation module.
所述获取模块用于获取一头部目标的三张二维影像，分别为第一影像、第二影像以及第三影像，其中，所述第一影像、第三影像与第二影像的拍摄方向的夹角取值范围均为[75，105]；The acquisition module is used to acquire three two-dimensional images of a head target, which are a first image, a second image and a third image, wherein the angles between the shooting directions of the first image, the third image and the second image are in the range of [75, 105];
所述计算模块用于通过人工智能算法利用所述三张二维影像获取所述头部目标的三维模型。The computing module is used to obtain a three-dimensional model of the head target using the three two-dimensional images through an artificial intelligence algorithm.
较佳地，所述数据处理组件还包括一数据模块、一整理模块以及一训练模块，Preferably, the data processing component further includes a data module, a sorting module and a training module.
所述数据模块用于获取一三维模型数据库，所述三维模型数据库包括原始三维影像；The data module is used to obtain a three-dimensional model database, and the three-dimensional model database includes original three-dimensional images;
所述整理模块用于对每一原始三维影像做标准化处理获取标准三维影像；The sorting module is used to perform standardization processing on each original three-dimensional image to obtain a standard three-dimensional image;
所述训练模块用于利用全部所述标准三维模型做人工智能学习获取训练模板；The training module is used to use all the standard three-dimensional models to perform artificial intelligence learning to obtain training templates;
所述计算模块用于通过所述训练模板利用所述三张二维影像获取所述头部目标的三维模型。The calculation module is used to obtain the three-dimensional model of the head target using the three two-dimensional images through the training template.
较佳地，所述数据处理组件还包括一识别模块、一处理模块以及一匹配模块，Preferably, the data processing component further includes an identification module, a processing module and a matching module.
所述识别模块用于识别第一影像、第二影像以及第三影像的轮廓特征；The recognition module is used to recognize contour features of the first image, the second image, and the third image;
所述处理模块用于根据所述轮廓特征获取二维影像的拍摄方向；The processing module is used to obtain the shooting direction of the two-dimensional image according to the contour feature;
所述匹配模块用于按所述拍摄方向查找与所述轮廓特征匹配的训练模板；The matching module is used to search for a training template matching the contour feature according to the shooting direction;
所述计算模块用于根据所述轮廓特征调整所述训练模板的形状以获取所述头部目标的三维模型。The calculation module is used to adjust the shape of the training template according to the contour feature to obtain a three-dimensional model of the head target.
在符合本领域常识的基础上，上述各优选条件，可任意组合，即得本发明各较佳实例。On the basis of being in accordance with the common sense in the art, the above-mentioned preferred conditions can be arbitrarily combined to obtain the preferred embodiments of the present invention.
本发明的积极进步效果在于：The positive and progressive effects of the present invention are:
本发明的基于二维影像生成三维模型的数据处理方法及组件能够利用三张二维影像生成3D模型，以能够获取高画质的3D模型。另外与现有的人脸识别系统结合，大幅度提高大角度时人脸识别的精度，，可支持多达90度完全侧脸时的人脸识别，在摄像头过高、在拍摄目标偏转头部时依然能快速锁定目标。The data processing method and component for generating a three-dimensional model based on two-dimensional images of the present invention can generate a 3D model using three two-dimensional images to obtain a high-quality 3D model. In addition, combined with the existing face recognition system, the accuracy of face recognition at large angles is greatly improved, and face recognition at up to 90 degrees of complete side faces can be supported. When the camera is too high or the target is tilted, the target can still be quickly locked.
附图说明BRIEF DESCRIPTION OF THE DRAWINGS
图1为本发明实施例1的数据处理方法的流程图。FIG1 is a flow chart of a data processing method according to Embodiment 1 of the present invention.
图2为本发明实施例1的数据处理方法的另一流程图。FIG. 2 is another flow chart of the data processing method according to the first embodiment of the present invention.
图3为本发明实施例1的数据处理方法的又一流程图。FIG. 3 is another flow chart of the data processing method according to the first embodiment of the present invention.
图4为本发明实施例1的数据处理方法的又一流程图。FIG. 4 is another flow chart of the data processing method according to the first embodiment of the present invention.
具体实施方式Detailed ways
下面通过实施例的方式进一步说明本发明，但并不因此将本发明限制在所述的实施例范围之中。The present invention is further described below by way of examples, but the present invention is not limited to the scope of the examples.
实施例1Example 1
本实施例提供一种基于二维影像生成三维模型的数据处理组件。This embodiment provides a data processing component for generating a three-dimensional model based on a two-dimensional image.
所述数据处理组件包括一获取模块以及一计算模块，The data processing component includes an acquisition module and a calculation module.
所述获取模块用于获取一头部目标的三张二维影像，分别为第一影像、第二影像以及第三影像，其中，所述第一影像、第三影像与第二影像的拍摄方向的夹角取值范围均为[75，105]；The acquisition module is used to acquire three two-dimensional images of a head target, which are a first image, a second image and a third image, wherein the angles between the shooting directions of the first image, the third image and the second image are in the range of [75, 105];
在本实施例中所述夹角的取值范围为[85，95]，最优拍摄方向的夹角取值为90度。In this embodiment, the angle ranges from [85, 95], and the angle of the optimal shooting direction is 90 degrees.
所述计算模块用于通过人工智能算法利用所述三张二维影像获取所述头部目标的三维模型。The computing module is used to obtain a three-dimensional model of the head target using the three two-dimensional images through an artificial intelligence algorithm.
具体地，所述数据处理组件还包括一数据模块、一整理模块以及一训练模块，上述三个模块用于实现通过人工智能算法利用所述三张二维影像获取所述头部目标的三维模型。Specifically, the data processing component also includes a data module, a sorting module and a training module. The above three modules are used to obtain the three-dimensional model of the head target using the three two-dimensional images through an artificial intelligence algorithm.
所述数据模块用于获取一三维模型数据库，所述三维模型数据库包括原始三维影像；The data module is used to obtain a three-dimensional model database, and the three-dimensional model database includes original three-dimensional images;
所述整理模块用于对每一原始三维影像做标准化处理获取标准三维影像；The sorting module is used to perform standardization processing on each original three-dimensional image to obtain a standard three-dimensional image;
所述训练模块用于利用全部所述标准三维模型做人工智能学习获取训练模板；The training module is used to use all the standard three-dimensional models to perform artificial intelligence learning to obtain training templates;
所述计算模块用于通过所述训练模板利用所述三张二维影像获取所述头部目标的三维模型。The calculation module is used to obtain the three-dimensional model of the head target using the three two-dimensional images through the training template.
本实施例的所述数据处理组件还包括一截取模块。The data processing component of this embodiment also includes an interception module.
对于每一标准三维模型，所述截取模块用于获取所述标准三维模型的三张二维截图，分别为第一截图、第二截图以及第三截图，其中，所述第一截图、第三截图与第二截图的截取方向的夹角取值范围均为[75，105]；For each standard three-dimensional model, the interception module is used to obtain three two-dimensional screenshots of the standard three-dimensional model, namely a first screenshot, a second screenshot and a third screenshot, wherein the angles between the interception directions of the first screenshot, the third screenshot and the second screenshot are all in the range of [75, 105];
本实施例中的截取是将标准三维模型放置在一目标平面的一侧，所述目标平面与截取方向也就是观测方向(等同于拍摄方向)垂直，然后将标准三维模型上的像素点由近到远依次垂直落在所述目标平面上，生成所述二维截图。The interception in this embodiment is to place the standard three-dimensional model on one side of a target plane, and the target plane is perpendicular to the interception direction, that is, the observation direction (equivalent to the shooting direction), and then the pixel points on the standard three-dimensional model are vertically dropped on the target plane from near to far in sequence to generate the two-dimensional screenshot.
所述训练模块用于利用全部所述标准三维模型做人工智能学习获取训练模板，所述训练模板包括三张二维截图与标准三维模型的对应关系。The training module is used to use all the standard three-dimensional models to perform artificial intelligence learning to obtain a training template, and the training template includes the correspondence between three two-dimensional screenshots and the standard three-dimensional model.
所述数据处理组件还包括一识别模块、一处理模块以及一匹配模块，The data processing component also includes an identification module, a processing module and a matching module.
所述识别模块用于识别第一影像、第二影像以及第三影像的轮廓特征；The recognition module is used to recognize contour features of the first image, the second image, and the third image;
所述处理模块用于根据所述轮廓特征获取二维影像的拍摄方向；The processing module is used to obtain the shooting direction of the two-dimensional image according to the contour feature;
所述匹配模块用于按所述拍摄方向查找与所述轮廓特征匹配的训练模板；The matching module is used to search for a training template matching the contour feature according to the shooting direction;
所述计算模块用于根据所述轮廓特征调整所述训练模板的形状以获取所述头部目标的三维模型。The calculation module is used to adjust the shape of the training template according to the contour feature to obtain a three-dimensional model of the head target.
进一步地，本实施例的头部目标的三维模型包括结构层和像素层，所述计算模块用于根据所述轮廓特征调整所述训练模板的形状以获取三维模型的结构层，在获取结构层之后，沿轮廓特征获取二维影像的拍摄方向在所述结构层上贴设头部目标的三张二维影像以获取所述三维模型。Furthermore, the three-dimensional model of the head target in this embodiment includes a structural layer and a pixel layer. The calculation module is used to adjust the shape of the training template according to the contour feature to obtain the structural layer of the three-dimensional model. After obtaining the structural layer, three two-dimensional images of the head target are pasted on the structural layer along the shooting direction of the two-dimensional image obtained along the contour feature to obtain the three-dimensional model.
所述数据处理方法包括一标准化组件，所述标准化组件为张量模型。The data processing method includes a standardization component, which is a tensor model.
所述张量模型可以为预存影像上设置的表示影像点之间关系的函数式，所述生成模块用于通过人工智能深度学习算法利用所述预存影像上函数式来设置所述3D影像上影像点之间的函数式。The tensor model may be a function formula representing the relationship between image points set on a pre-stored image, and the generation module is used to set the function formula between image points on the 3D image using an artificial intelligence deep learning algorithm using the function formula on the pre-stored image.
对于每一原始三维影像，所述整理模块用于按所述原始三维模型调节所述标准化模型以获取所述标准三维影像。For each original 3D image, the sorting module is used to adjust the standardized model according to the original 3D model to obtain the standard 3D image.
具体地，所述整理模块用于：Specifically, the collating module is used for:
识别原始三维影像的五官特征点；Identify facial feature points of the original 3D image;
根据所述五官特征点调节所述标准化模型的尺寸，并将原始三维影像与标准化模型在空间中对齐；Adjusting the size of the standardized model according to the facial feature points, and aligning the original three-dimensional image with the standardized model in space;
对于标准化模型上的每一影像点，获取影像点处法线与原始三维影像的交点；For each image point on the standardized model, obtain the intersection point of the normal line at the image point and the original three-dimensional image;
按预设规则调节所述标准化模型以获取所述标准三维影像，所述预设规则为将每一影像点与对应交点的之间的长度调节为相同长度。The standardized model is adjusted according to a preset rule to obtain the standard three-dimensional image, and the preset rule is to adjust the length between each image point and the corresponding intersection point to the same length.
参见图1，利用上述数据处理组件，本实施例还提供一种数据处理方法，包括：Referring to FIG. 1 , using the above data processing component, this embodiment further provides a data processing method, including:
步骤101、获取一头部目标的三张二维影像，分别为第一影像、第二影像以及第三影像，其中，所述第一影像、第三影像与第二影像的拍摄方向的夹角取值范围均为[75，105]；Step 101: Acquire three two-dimensional images of a head target, which are a first image, a second image, and a third image, wherein the angles between the shooting directions of the first image, the third image, and the second image are all in the range of [75, 105];
在本实施例中所述夹角的取值范围为[85，95]，最优拍摄方向的夹角取值为90度。In this embodiment, the angle ranges from [85, 95], and the angle of the optimal shooting direction is 90 degrees.
步骤102、通过人工智能算法利用所述三张二维影像获取所述头部目标的三维模型。Step 102: Acquire a three-dimensional model of the head target using the three two-dimensional images through an artificial intelligence algorithm.
在步骤101之前，所述数据处理方法还包括：Before step 101, the data processing method further includes:
步骤1001、获取一三维模型数据库，所述三维模型数据库包括原始三维影像；Step 1001: Acquire a 3D model database, wherein the 3D model database includes an original 3D image;
步骤1002、对每一原始三维影像做标准化处理获取标准三维影像；Step 1002: Perform standardization processing on each original 3D image to obtain a standard 3D image;
步骤1003、利用全部所述标准三维模型做人工智能学习获取训练模板；Step 1003: Utilize all the standard three-dimensional models to perform artificial intelligence learning to obtain training templates;
步骤102为：通过所述训练模板利用所述三张二维影像获取所述头部目标的三维模型。也就是说步骤102具体为：通过人工智能算法以及所述训练模板利用所述三张二维影像获取所述头部目标的三维模型。Step 102 is: using the three two-dimensional images through the training template to obtain the three-dimensional model of the head target. That is to say, step 102 specifically is: using the three two-dimensional images through the artificial intelligence algorithm and the training template to obtain the three-dimensional model of the head target.
参见图2，步骤1003包括：Referring to FIG. 2 , step 1003 includes:
步骤10031、对于每一标准三维模型，获取所述标准三维模型数据特征表达，并对所述数据特征表达进行统计以得到数据特征表达的平均数据还有均方差数据；Step 10031: for each standard three-dimensional model, obtain the data feature expression of the standard three-dimensional model, and perform statistics on the data feature expression to obtain average data and mean square error data of the data feature expression;
步骤10032、利用全部所述标准三维模型的所述平均数据和所述均方差数据做人工智能学习获取训练模板，所述训练模板通过控制参数以重建出不同人脸模型。Step 10032: Utilize the average data and the mean square error data of all the standard three-dimensional models to perform artificial intelligence learning to obtain a training template, and the training template is used to reconstruct different face models by controlling parameters.
参见图3，步骤102包括：Referring to FIG. 3 , step 102 includes:
步骤1021、识别第一影像、第二影像以及第三影像的轮廓特征；Step 1021, identifying contour features of the first image, the second image, and the third image;
步骤1022、根据所述轮廓特征获取二维影像的拍摄方向；Step 1022, obtaining a shooting direction of the two-dimensional image according to the contour feature;
步骤1023、按所述拍摄方向查找与所述轮廓特征匹配的训练模板；Step 1023, searching for a training template matching the contour feature according to the shooting direction;
步骤1024、根据所述轮廓特征调整所述训练模板的形状以获取所述头部目标的三维模型。Step 1024: Adjust the shape of the training template according to the contour feature to obtain a three-dimensional model of the head target.
所述数据处理方法包括一标准化模型，所述标准化模型为张量模型，步骤1002包括：The data processing method includes a standardized model, which is a tensor model. Step 1002 includes:
对于每一原始三维影像，按所述原始三维模型调节所述标准化模型以获取所述标准三维影像。For each original 3D image, the standardized model is adjusted according to the original 3D model to obtain the standardized 3D image.
参见图4，具体地，步骤1002包括：Referring to FIG. 4 , specifically, step 1002 includes:
10021、识别原始三维影像的五官特征点；10021. Identify facial feature points of the original three-dimensional image;
10022、根据所述五官特征点调节所述标准化模型的尺寸，并将原始三维影像与标准化模型在空间中对齐；10022. Adjust the size of the standardized model according to the facial feature points, and align the original three-dimensional image with the standardized model in space;
10023、对于标准化模型上的每一影像点，获取影像点处法线与原始三维影像的交点；10023. For each image point on the standardized model, obtain the intersection point of the normal line at the image point and the original three-dimensional image;
进一步地，步骤10023中还可以包括，在预设区域(如脸部、额头等)获取影像点处法线与相邻影像点法线的夹角，若夹角小于预设值则获取影像点处法线与原始三维影像的交点，若所述夹角大于另一预设值则将所述影像点作为噪点，做平整化处理，处理的方式可以为获取噪点影像点周围指定区域中法线夹角的平均值，然后按所述平均值设置所述噪点影像点。Furthermore, step 10023 may also include obtaining the angle between the normal at the image point and the normal of the adjacent image point in a preset area (such as the face, forehead, etc.); if the angle is less than a preset value, obtaining the intersection of the normal at the image point and the original three-dimensional image; if the angle is greater than another preset value, treating the image point as a noise point and performing smoothing processing; the processing method may be to obtain the average value of the normal angles in a specified area around the noise image point, and then setting the noise image point according to the average value.
10024、按预设规则调节所述标准化模型以获取所述标准三维影像，所述预设规则为将每一影像点与对应交点的之间的长度调节为相同长度。10024. Adjust the standardized model according to a preset rule to obtain the standard three-dimensional image, wherein the preset rule is to adjust the length between each image point and the corresponding intersection point to the same length.
本实施例的基于二维影像生成三维模型的数据处理方法及组件能够利用三张二维影像生成3D模型，提高人脸识别的精度，且能够获取高画质的3D模型，产品支持多达90度完全侧脸时的人脸识别，在摄像头过高、在拍摄目标偏转头部时依然能快速锁定目标。The data processing method and components for generating three-dimensional models based on two-dimensional images in this embodiment can generate a 3D model using three two-dimensional images, thereby improving the accuracy of face recognition and obtaining a high-quality 3D model. The product supports face recognition up to 90 degrees of complete side face, and can quickly lock onto the target even when the camera is too high or the target's head is turned.
虽然以上描述了本发明的具体实施方式，但是本领域的技术人员应当理解，这些仅是举例说明，本发明的保护范围是由所附权利要求书限定的。本领域的技术人员在不背离本发明的原理和实质的前提下，可以对这些实施方式做出多种变更或修改，但这些变更和修改均落入本发明的保护范围。Although the specific embodiments of the present invention are described above, it should be understood by those skilled in the art that these are only examples, and the protection scope of the present invention is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but these changes and modifications all fall within the protection scope of the present invention.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911258760.1ACN111161399B (en) | 2019-12-10 | 2019-12-10 | Data processing method and component for generating three-dimensional model based on two-dimensional image |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911258760.1ACN111161399B (en) | 2019-12-10 | 2019-12-10 | Data processing method and component for generating three-dimensional model based on two-dimensional image |
| Publication Number | Publication Date |
|---|---|
| CN111161399A CN111161399A (en) | 2020-05-15 |
| CN111161399Btrue CN111161399B (en) | 2024-04-19 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911258760.1AActiveCN111161399B (en) | 2019-12-10 | 2019-12-10 | Data processing method and component for generating three-dimensional model based on two-dimensional image |
| Country | Link |
|---|---|
| CN (1) | CN111161399B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111882656A (en)* | 2020-06-19 | 2020-11-03 | 深圳宏芯宇电子股份有限公司 | Graph processing method, equipment and storage medium based on artificial intelligence |
| CN113888692B (en)* | 2020-12-11 | 2024-09-06 | 深圳钒佳科技有限公司 | System and method for converting two-dimensional image into three-dimensional image by deep learning |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006107145A (en)* | 2004-10-05 | 2006-04-20 | Tokyo Univ Of Agriculture & Technology | Face shape modeling system and face shape modeling method |
| WO2015161728A1 (en)* | 2014-04-22 | 2015-10-29 | 重庆海扶医疗科技股份有限公司 | Three-dimensional model construction method and device, and image monitoring method and device |
| CN109377544A (en)* | 2018-11-30 | 2019-02-22 | 腾讯科技（深圳）有限公司 | A kind of face three-dimensional image generating method, device and readable medium |
| CN109636926A (en)* | 2018-11-23 | 2019-04-16 | 盎锐（上海）信息科技有限公司 | 3D overall situation Free Transform algorithm and device |
| CN110019901A (en)* | 2017-09-13 | 2019-07-16 | 深圳三维盘酷网络科技有限公司 | Three-dimensional model search device, searching system, search method and computer readable storage medium |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006107145A (en)* | 2004-10-05 | 2006-04-20 | Tokyo Univ Of Agriculture & Technology | Face shape modeling system and face shape modeling method |
| WO2015161728A1 (en)* | 2014-04-22 | 2015-10-29 | 重庆海扶医疗科技股份有限公司 | Three-dimensional model construction method and device, and image monitoring method and device |
| CN110019901A (en)* | 2017-09-13 | 2019-07-16 | 深圳三维盘酷网络科技有限公司 | Three-dimensional model search device, searching system, search method and computer readable storage medium |
| CN109636926A (en)* | 2018-11-23 | 2019-04-16 | 盎锐（上海）信息科技有限公司 | 3D overall situation Free Transform algorithm and device |
| CN109377544A (en)* | 2018-11-30 | 2019-02-22 | 腾讯科技（深圳）有限公司 | A kind of face three-dimensional image generating method, device and readable medium |
| Title |
|---|
| 胡向阳 ; 侯文广 ; 丁明跃 ; .基于真实三维信息的人脸自动识别.中国人民公安大学学报(自然科学版).2008,(03),全文.* |
| Publication number | Publication date |
|---|---|
| CN111161399A (en) | 2020-05-15 |
| Publication | Publication Date | Title |
|---|---|---|
| US11632537B2 (en) | Method and apparatus for obtaining binocular panoramic image, and storage medium | |
| US10460521B2 (en) | Transition between binocular and monocular views | |
| US10380421B2 (en) | Iris recognition via plenoptic imaging | |
| CN109558764A (en) | Face identification method and device, computer equipment | |
| CN109064397A (en) | A kind of image split-joint method and system based on camera shooting earphone | |
| KR101510312B1 (en) | 3D face-modeling device, system and method using Multiple cameras | |
| CN105427385A (en) | High-fidelity face three-dimensional reconstruction method based on multilevel deformation model | |
| CN107656619A (en) | A kind of intelligent projecting method, system and intelligent terminal | |
| JP7479729B2 (en) | Three-dimensional representation method and device | |
| CN108600729B (en) | Dynamic 3D model generation device and image generation method | |
| CN110909634A (en) | Visible light and double infrared combined rapid in vivo detection method | |
| CN110245549A (en) | Real-time face and object manipulation | |
| CN111161399B (en) | Data processing method and component for generating three-dimensional model based on two-dimensional image | |
| CN108347505B (en) | Mobile terminal with 3D imaging function and image generation method | |
| CN111402395A (en) | A Passive Polarization 3D Reconstruction Method Based on CNN Correction | |
| CN119169288A (en) | A scene object classification method and system based on multi-view depth images | |
| CN108391116B (en) | Whole body scanning device and method based on 3D imaging technology | |
| CN111145361A (en) | Naked eye 3D display vision improving method | |
| CN107592491B (en) | Video communication background display method and device | |
| Li et al. | Learning to synthesize photorealistic dual-pixel images from RGBD frames | |
| CN109360270B (en) | 3D face pose alignment method and device based on artificial intelligence | |
| CN104751144A (en) | Frontal face quick evaluation method for video surveillance | |
| CN109151440B (en) | Image positioning device and method | |
| CN109657559B (en) | Point cloud depth perception coding engine device | |
| CN108921102B (en) | 3D image processing method and device |
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information | Address after:518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Applicant after:Angwei Cloud (Shenzhen) Computing Co.,Ltd. Address before:518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Applicant before:Angrui (Shenzhen) Information Technology Co.,Ltd. | |
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right | Effective date of registration:20230807 Address after:201703 Room 2134, Floor 2, No. 152 and 153, Lane 3938, Huqingping Road, Qingpu District, Shanghai Applicant after:Shanghai Qingyan Heshi Technology Co.,Ltd. Address before:518000 Room 201, building A, No. 1, Qian Wan Road, Qianhai Shenzhen Hong Kong cooperation zone, Shenzhen, Guangdong (Shenzhen Qianhai business secretary Co., Ltd.) Applicant before:Angwei Cloud (Shenzhen) Computing Co.,Ltd. | |
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right | ||
| PE01 | Entry into force of the registration of the contract for pledge of patent right | Denomination of invention:Data processing method and components for generating 3D models based on 2D images Granted publication date:20240419 Pledgee:Shanghai Rural Commercial Bank Co.,Ltd. Qingpu sub branch Pledgor:Shanghai Qingyan Heshi Technology Co.,Ltd. Registration number:Y2025310000521 |