技术领域technical field
本公开涉及人脸识别技术领域，尤其涉及一种人脸识别方法和装置。The present disclosure relates to the technical field of face recognition, and in particular to a face recognition method and device.
背景技术Background technique
人脸是人们相互判别、认识、记忆的主要标识，人脸识别在计算机视觉、模式识别、多媒体技术研究中占有重要的地位。Face is the main symbol for people to distinguish, recognize and remember each other. Face recognition plays an important role in the research of computer vision, pattern recognition and multimedia technology.
相关技术中，人脸识别的方法一般都是依次两幅人脸图像进行人脸检测、特征点定位、特征提取，并根据提取的特征进行相似性度量，得到用于衡量两幅人脸图像相似度的分数。In related technologies, the method of face recognition is generally to perform face detection, feature point location, and feature extraction on two face images in sequence, and perform similarity measurement according to the extracted features, and obtain a measure of the similarity of two face images. degree score.
当两幅人脸图像中人脸上均有深框眼镜，或者两幅人脸图像中人脸具有相同或相似的发型时，本来相似度不是很高的两个人脸，很有可能会被认为相似度较高，因此识别的准确率较低。When there are deep-frame glasses on the faces of the two face images, or the faces in the two face images have the same or similar hairstyles, the two faces whose similarity is not very high are likely to be regarded as The similarity is higher, so the recognition accuracy is lower.
发明内容Contents of the invention
为了克服相关技术中存在的识别的准确率较低的问题，本公开提供一种人脸识别方法和装置。所述技术方案如下：In order to overcome the problem of low recognition accuracy in related technologies, the present disclosure provides a face recognition method and device. Described technical scheme is as follows:
根据本公开实施例的第一方面，提供一种人脸识别方法，适用于判断两个人脸图像的相似度，包括：According to the first aspect of an embodiment of the present disclosure, a face recognition method is provided, which is suitable for judging the similarity of two face images, including:
获取第一人脸图像；Obtain the first face image;
确定所述第一人脸图像和指定人脸图像的相似度；determining the similarity between the first face image and the designated face image;
确定所述第一人脸图像和所述指定人脸图像的干扰特征对所述相似度的干扰值；Determining the interference value of the interference features of the first human face image and the designated human face image on the similarity;
根据所述干扰值，调整所述相似度。Adjust the similarity according to the interference value.
在本公开的第一种可能的实现方式中，所述确定所述第一人脸图像和所述指定人脸图像的干扰特征对所述相似度的干扰值，包括：In a first possible implementation manner of the present disclosure, the determining the interference value of the similarity between the interference features of the first face image and the specified face image includes:
将所述第一人脸图像和所述指定人脸图像分别与设定的平均形状模型对齐；Aligning the first human face image and the specified human face image with the set average shape model respectively;
分别对对齐后的所述第一人脸图像和对齐后的所述指定人脸图像进行肤色分析，确定对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域；Perform skin color analysis on the aligned first human face image and the aligned designated human face image respectively, and determine the non-skin color of the aligned first human face image and the aligned designated human face image area;
根据所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域，计算所述第一人脸图像和所述指定人脸图像的干扰特征对相似度的干扰值。According to the aligned first human face image and the non-skin color area of the aligned specified human face image, calculate the interference feature pair similarity between the first human face image and the specified human face image Interference value.
在本公开的第二种可能的实现方式中，所述分别对对齐后的所述第一人脸图像和对齐后的所述指定人脸图像进行肤色分析，确定对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域，包括：In a second possible implementation manner of the present disclosure, the skin color analysis is performed on the aligned first human face image and the aligned specified human face image, and the aligned first human face image is determined The face image and the non-skin color area of the specified face image after alignment include:
选取对齐后的所述第一人脸图像和对齐后的所述指定人脸图像中，与所述平均形状人脸模型中的设定肤色区域对应的区域为第一肤色区域；Selecting the aligned first human face image and the aligned designated human face image, the area corresponding to the set skin color area in the average shape face model is the first skin color area;
提取所述第一肤色区域的肤色特征，并将对齐后的所述第一人脸图像和对齐后的所述指定人脸图像中，肤色特征与所述第一肤色区域的肤色特征相同的区域确定为第二肤色区域；extracting the skin color feature of the first skin color area, and aligning the first face image after alignment and the designated face image after alignment, the area with the same skin color feature as the skin color feature of the first skin color area Determined as the second skin color area;
将对齐后的所述第一人脸图像和对齐后的所述指定人脸图像中，除所述第一肤色区域和所述第二肤色区域以外的所有区域，作为所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域。In the aligned first human face image and the aligned specified human face image, all areas except the first skin color area and the second skin color area are used as the aligned The first human face image and the aligned non-skin color region of the specified human face image.
在本公开的第三种可能的实现方式中，所述根据所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域，计算所述第一人脸图像和所述指定人脸图像的干扰特征对相似度的干扰值，包括：In a third possible implementation manner of the present disclosure, the calculation of the first human face is based on the aligned first human face image and the aligned non-skin color area of the specified human face image. The interference value of the interference feature of the image and the specified face image to the similarity includes:
分别将所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域的像素的特征值取为1，所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的肤色区域的特征值取为0；The eigenvalues of the pixels of the non-skin color region of the aligned first human face image and the aligned designated human face image are respectively taken as 1, and the aligned first human face image and The eigenvalue of the skin color region of the specified face image after alignment is taken as 0;
对对齐后的所述第一人脸图像和对齐后的所述指定人脸图像进行图像交；performing image intersection on the aligned first human face image and the aligned specified human face image;
统计对齐后的所述第一人脸图像和对齐后的所述指定人脸图像对应的像素的特征值均为1的像素的数量；Counting the number of pixels whose eigenvalues of the pixels corresponding to the aligned first face image and the aligned specified face image are all 1;
计算统计得到的数量与对齐后的所述第一人脸图像或对齐后的所述指定人脸图像的像素的总数的比值，得到所述第一人脸图像和所述指定人脸图像的干扰特征对相似度的干扰值。Calculate the ratio of the number obtained by statistics to the total number of pixels of the aligned first human face image or the aligned designated human face image to obtain the interference between the first human face image and the designated human face image The noise value of the feature on the similarity.
在本公开的第四种可能的实现方式中，所述根据所述干扰值，调整所述相似度，包括：In a fourth possible implementation manner of the present disclosure, the adjusting the similarity according to the interference value includes:
按照预定的函数关系，根据所述干扰值，确定所述相似度的修正值；According to a predetermined functional relationship, according to the interference value, determine the correction value of the similarity;
将所述相似度减去所述修正值，得到调整后的所述相似度。The correction value is subtracted from the similarity to obtain the adjusted similarity.
根据本公开实施例的第二方面，提供一种人脸识别装置，适用于判断两个人脸图像的相似度，包括：According to the second aspect of the embodiments of the present disclosure, a face recognition device is provided, which is suitable for judging the similarity of two face images, including:
获取模块，用于获取第一人脸图像；An acquisition module, configured to acquire the first face image;
识别模块，用于确定所述第一人脸图像和指定人脸图像的相似度；A recognition module, configured to determine the similarity between the first face image and a designated face image;
干扰确定模块，用于确定所述第一人脸图像和所述指定人脸图像的干扰特征对所述相似度的干扰值；An interference determination module, configured to determine an interference value of the similarity between the interference features of the first face image and the designated face image;
修正模块，用于根据所述干扰值，调整所述相似度。A correction module, configured to adjust the similarity according to the interference value.
在本公开的第一种可能的实现方式中，所述干扰确定模块包括：In a first possible implementation manner of the present disclosure, the interference determination module includes:
对齐单元，用于将所述第一人脸图像和所述指定人脸图像分别与设定的平均形状模型对齐；an alignment unit, configured to align the first face image and the designated face image with a set average shape model;
区域确定单元，用于分别对对齐后的所述第一人脸图像和对齐后的所述指定人脸图像进行肤色分析，确定对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域；An area determining unit, configured to perform skin color analysis on the aligned first human face image and the aligned specified human face image, and determine the aligned first human face image and the aligned specified human face image. The non-skinned area of the face image;
干扰计算单元，用于根据所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域，计算所述第一人脸图像和所述指定人脸图像的干扰特征对相似度的干扰值。An interference calculation unit, configured to calculate the difference between the first human face image and the specified human face image according to the aligned non-skin color area of the first human face image and the aligned specified human face image The interference value of the interference feature on the similarity.
在本公开的第二种可能的实现方式中，所述区域确定单元用于，In a second possible implementation manner of the present disclosure, the area determination unit is configured to:
选取对齐后的所述第一人脸图像和对齐后的所述指定人脸图像中，与所述平均形状人脸模型中的设定肤色区域对应的区域为第一肤色区域；Selecting the aligned first human face image and the aligned designated human face image, the area corresponding to the set skin color area in the average shape face model is the first skin color area;
提取所述第一肤色区域的肤色特征，并将对齐后的所述第一人脸图像和对齐后的所述指定人脸图像中，肤色特征与所述第一肤色区域的肤色特征相同的区域确定为第二肤色区域；extracting the skin color feature of the first skin color area, and aligning the first face image after alignment and the designated face image after alignment, the area with the same skin color feature as the skin color feature of the first skin color area Determined as the second skin color area;
将对齐后的所述第一人脸图像和对齐后的所述指定人脸图像中，除所述第一肤色区域和所述第二肤色区域以外的所有区域，作为所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域。In the aligned first human face image and the aligned specified human face image, all areas except the first skin color area and the second skin color area are used as the aligned The first human face image and the aligned non-skin color region of the specified human face image.
在本公开的第三种可能的实现方式中，所述干扰计算单元用于，In a third possible implementation manner of the present disclosure, the interference calculation unit is configured to:
分别将所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的非肤色区域的像素的特征值取为1，所述对齐后的所述第一人脸图像和对齐后的所述指定人脸图像的肤色区域的特征值取为0；The eigenvalues of the pixels of the non-skin color region of the aligned first human face image and the aligned designated human face image are respectively taken as 1, and the aligned first human face image and The eigenvalue of the skin color region of the specified face image after alignment is taken as 0;
对对齐后的所述第一人脸图像和对齐后的所述指定人脸图像进行图像交；performing image intersection on the aligned first human face image and the aligned specified human face image;
统计对齐后的所述第一人脸图像和对齐后的所述指定人脸图像对应的像素的特征值均为1的像素的数量；Counting the number of pixels whose eigenvalues of the pixels corresponding to the aligned first face image and the aligned specified face image are all 1;
计算统计得到的数量与对齐后的所述第一人脸图像或对齐后的所述指定人脸图像的像素的总数的比值，得到所述第一人脸图像和所述指定人脸图像的干扰特征对相似度的干扰值。Calculate the ratio of the number obtained by statistics to the total number of pixels of the aligned first human face image or the aligned designated human face image to obtain the interference between the first human face image and the designated human face image The noise value of the feature on the similarity.
在本公开的第四种可能的实现方式中，所述修正模块包括：In a fourth possible implementation manner of the present disclosure, the correction module includes:
修正值确定单元，用于按照预定的函数关系，根据所述干扰值，确定所述相似度的修正值；A correction value determination unit, configured to determine a correction value of the similarity according to the interference value according to a predetermined functional relationship;
分数计算单元，用于将所述相似度减去所述修正值，得到调整后的所述相似度。A score calculation unit, configured to subtract the correction value from the similarity to obtain the adjusted similarity.
根据本公开实施例的第三方面，提供一种人脸识别装置，适用于判断两个人脸图像的相似度，包括：According to a third aspect of an embodiment of the present disclosure, a face recognition device is provided, which is suitable for judging the similarity of two face images, including:
处理器；processor;
用于存储处理器可执行指令的存储器；memory for storing processor-executable instructions;
其中，所述处理器被配置为：Wherein, the processor is configured as:
获取第一人脸图像；Obtain the first face image;
确定所述第一人脸图像和指定人脸图像的相似度；determining the similarity between the first face image and the designated face image;
确定所述第一人脸图像和所述指定人脸图像的干扰特征对所述相似度的干扰值；Determining the interference value of the interference features of the first human face image and the designated human face image on the similarity;
根据所述干扰值，调整所述相似度。Adjust the similarity according to the interference value.
本公开的实施例提供的技术方案可以包括以下有益效果：通过确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值，并根据该干扰值调整第一人脸图像和指定人脸图像的相似度，避免由于第一人脸图像和指定人脸图像中人脸上均有深框眼镜，或者第一人脸图像和指定人脸图像中人脸具有相同或相似的发型等原因，造成的将相似度不是很高的两个人脸误判为相似度较高，提高了识别的准确率。The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects: by determining the interference value of the interference features of the first face image and the specified face image to the similarity, and adjusting the first face image and the specified face image according to the interference value The similarity of the face images, avoiding that there are deep-frame glasses on the faces of the first face image and the designated face image, or the faces in the first face image and the designated face image have the same or similar hairstyle, etc. The reason is that two faces whose similarity is not very high are misjudged as having a high similarity, which improves the accuracy of recognition.
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，并不能限制本公开。It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure.
附图说明Description of drawings
此处的附图被并入说明书中并构成本说明书的一部分，示出了符合本发明的实施例，并与说明书一起用于解释本发明的原理。The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description serve to explain the principles of the invention.
图1是根据一示例性实施例示出的一种人脸识别方法的流程图；Fig. 1 is a flowchart of a face recognition method shown according to an exemplary embodiment;
图2是根据一示例性实施例示出的另一种人脸识别方法的流程图；Fig. 2 is a flowchart of another face recognition method shown according to an exemplary embodiment;
图3是根据一示例性实施例示出的一种人脸识别装置的框图；Fig. 3 is a block diagram of a face recognition device according to an exemplary embodiment;
图4是根据一示例性实施例示出的另一种人脸识别装置的框图；Fig. 4 is a block diagram of another face recognition device according to an exemplary embodiment;
图5是根据一示例性实施例示出的一种人脸识别装置的框图。Fig. 5 is a block diagram of a face recognition device according to an exemplary embodiment.
具体实施方式Detailed ways
这里将详细地对示例性实施例进行说明，其示例表示在附图中。下面的描述涉及附图时，除非另有表示，不同附图中的相同数字表示相同或相似的要素。以下示例性实施例中所描述的实施方式并不代表与本发明相一致的所有实施方式。相反，它们仅是与如所附权利要求书中所详述的、本发明的一些方面相一致的装置和方法的例子。Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numerals in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with aspects of the invention as recited in the appended claims.
图1是根据一示例性实施例示出的一种人脸识别方法的流程图，如图1所示，人脸识别方法用于移动终端中，适用于判断两个人脸图像的相似度，包括以下步骤。Fig. 1 is a flow chart of a face recognition method shown according to an exemplary embodiment. As shown in Fig. 1, the face recognition method is used in a mobile terminal and is suitable for judging the similarity between two face images, including the following step.
在步骤S101中，获取第一人脸图像。In step S101, a first face image is acquired.
在步骤S102中，确定第一人脸图像和指定人脸图像的相似度。In step S102, determine the similarity between the first face image and the designated face image.
在步骤S103中，确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。In step S103, determine the interference value of the similarity between the interference features of the first face image and the designated face image.
在本实施例中，干扰特征为第一人脸图像和指定人脸图像的人脸区域中特征值相同的非人脸部分的特征。干扰特征可以包括深框眼镜、头发、胡须等，本公开对此不作限制。干扰值用于衡量干扰特征对确定的相似度的干扰程度。In this embodiment, the interference feature is a feature of a non-face part having the same feature value in the face area of the first face image and the specified face image. Interfering features may include deep-frame glasses, hair, beard, etc., without limitation of the present disclosure. The noise value is used to measure how much the noise features interfere with the determined similarity.
在步骤S104中，根据干扰值，调整相似度。In step S104, the similarity is adjusted according to the interference value.
本公开实施例通过确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值，并根据该干扰值调整第一人脸图像和指定人脸图像的相似度，避免由于第一人脸图像和指定人脸图像中人脸上均有深框眼镜，或者第一人脸图像和指定人脸图像中人脸具有相同或相似的发型等原因，造成的将相似度不是很高的两个人脸误判为相似度较高，提高了识别的准确率。In the embodiment of the present disclosure, by determining the interference value of the similarity between the interference features of the first face image and the designated face image, and adjusting the similarity between the first face image and the designated face image according to the interference value, avoiding the interference caused by the first face image The similarity is not very high due to reasons such as the face image and the specified face image have deep-frame glasses on the face, or the faces in the first face image and the specified face image have the same or similar hairstyles, etc. The two faces are misjudged as having a higher similarity, which improves the accuracy of recognition.
图2是根据一示例性实施例示出的另一种人脸识别方法的流程图，如图2所示，人脸识别方法用于移动终端中，适用于判断两个人脸图像的相似度，包括以下步骤。Fig. 2 is a flow chart showing another face recognition method according to an exemplary embodiment. As shown in Fig. 2, the face recognition method is used in a mobile terminal and is suitable for judging the similarity between two face images, including The following steps.
在步骤S201中，获取第一人脸图像和第二人脸图像。In step S201, a first face image and a second face image are acquired.
在本实施例中，第二人脸图像为指定人脸图像，即指定人脸图像可以是从外界获取的。在其它实施例中，指定人脸图像也可以是预存在终端中的，本公开对此不作限制。In this embodiment, the second face image is a designated face image, that is, the designated face image may be obtained from the outside. In other embodiments, the designated face image may also be pre-stored in the terminal, which is not limited in the present disclosure.
在本实施例的一种实现方式中，该步骤S201可以包括：In an implementation manner of this embodiment, step S201 may include:
获取两幅图像；Get two images;
分别对两幅图像进行人脸检测，在两幅图像中的一幅图像中确定第一人脸图像，在两幅图像中的另一幅图像中确定第二人脸图像。Face detection is performed on the two images respectively, a first face image is determined in one of the two images, and a second face image is determined in the other image of the two images.
在实际应用中，对两幅图像进行人脸检测可以采用基于Adaboost(AdaptiveBoosting，自适应增强)的人脸检测算法。首先对图像按照预定的比例依次缩放，然后在每个图像的20*20像素的子窗口依次判别是人脸，还是非人脸，最后得到图像中人脸的位置和大小。根据图像中人脸的位置和大小，在两幅图像中进行截取，即可得到第一人脸图像和第二人脸图像。In practical applications, a face detection algorithm based on Adaboost (AdaptiveBoosting, adaptive enhancement) can be used for face detection of two images. First, the image is scaled sequentially according to a predetermined ratio, and then the 20*20 pixel sub-window of each image is sequentially judged whether it is a human face or a non-human face, and finally the position and size of the human face in the image are obtained. According to the position and size of the human face in the image, the first human face image and the second human face image can be obtained by intercepting the two images.
在步骤S202中，对第一人脸图像和第二人脸图像进行人脸识别，得到第一人脸图像和第二人脸图像的相似度。In step S202, face recognition is performed on the first face image and the second face image to obtain the similarity between the first face image and the second face image.
在本实施例的另一种实现方式中，该步骤S202可以包括：In another implementation manner of this embodiment, the step S202 may include:
分别采用基于ASM(Active Shape Model，主动形状模型)的特征点定位算法，确定第一人脸图像的形状模型和第二人脸图像的形状模型；Adopting the feature point localization algorithm based on ASM (Active Shape Model, Active Shape Model) respectively, determine the shape model of the first face image and the shape model of the second face image;
根据第一人脸图像的形状模型，对第一人脸图像依次进行Gabor(伽柏)小波变换、PCA(Principal Component Analysis，主成分分析)、LDA(LinearDiscriminant Analysis，线性判别分析)，得到第一人脸图像的特征信息；According to the shape model of the first face image, the first face image is sequentially subjected to Gabor wavelet transform, PCA (Principal Component Analysis, principal component analysis), LDA (Linear Discriminant Analysis, linear discriminant analysis), and the first Feature information of face image;
根据第二人脸图像的形状模型，对第二人脸图像进行Gabor小波变换、PCA、LDA，得到第二人脸图像的特征信息；According to the shape model of the second people's face image, Gabor wavelet transform, PCA, LDA are carried out to the second people's face image, obtain the feature information of the second people's face image;
计算第一人脸图像的特征信息和第二人脸图像的特征信息之间的余弦距离，并根据余弦距离，得到第一人脸图像和第二人脸图像的相似度。Calculate the cosine distance between the feature information of the first face image and the feature information of the second face image, and obtain the similarity between the first face image and the second face image according to the cosine distance.
在实际应用中，采用基于ASM的特征点定位算法确定人脸图像的形状模型时，先在图像中进行初始定位，再针对初始定位的各个特征点，根据每个特征点的灰度模型，在图像中搜索每个特征点的准确位置并进行修正。经过多次搜索和修正，确定的形状模型可以较好地反映人脸。In practical applications, when the feature point location algorithm based on ASM is used to determine the shape model of the face image, the initial location is performed in the image first, and then for each feature point of the initial location, according to the gray model of each feature point, in the The exact position of each feature point is searched in the image and corrected. After multiple searches and corrections, the determined shape model can better reflect the face.
第一人脸图像和第二人脸图像的相似度可以采用分数表示，如以100分为满分，90分表示第一人脸图像和第二人脸图像有90％的区域是相同的，相似度极高。The similarity between the first human face image and the second human face image can be represented by a score, such as 100 points for a full score, 90 points means that 90% of the first human face image and the second human face image have 90% of the same area, similar Extremely high.
可以理解地，通过执行步骤S202即可实现确定第一人脸图像和指定人脸图像的相似度。It can be understood that the determination of the similarity between the first human face image and the designated human face image can be realized by performing step S202.
在步骤S203中，将第一人脸图像和第二人脸图像分别与设定的平均形状模型对齐。In step S203, the first face image and the second face image are respectively aligned with the set average shape model.
在本实施例的又一种实现方式中，该步骤S203可以包括：In yet another implementation manner of this embodiment, the step S203 may include:
分别按照平均形状人脸模型，对第一人脸图像的形状模型和第二人脸图像的形状模型进行二维仿射变换。Perform two-dimensional affine transformation on the shape model of the first face image and the shape model of the second face image respectively according to the average shape face model.
在实际应用中，进行二维仿射变换时，只需针对人脸的形状模型和平均人脸形状模型计算转换函数即可。在对第一人脸图像的形状模型和第二人脸图像的形状模型进行二维仿射变换之后，可以得到大小相同的第一人脸图像和第二人脸图像，即与设定的平均形状模型对齐的第一人脸图像和第二人脸图像。In practical applications, when performing two-dimensional affine transformation, it is only necessary to calculate the conversion function for the shape model of the face and the average face shape model. After two-dimensional affine transformation is performed on the shape model of the first face image and the shape model of the second face image, the first face image and the second face image with the same size can be obtained, that is, the set average The shape model aligns the first face image and the second face image.
在步骤S204中，分别对对齐后的第一人脸图像和对齐后的第二人脸图像进行肤色分析，确定对齐后的第一人脸图像和对齐后的第二人脸图像的非肤色区域。In step S204, the skin color analysis is carried out to the aligned first human face image and the aligned second human face image respectively, and the non-skin color regions of the aligned first human face image and the aligned second human face image are determined .
在本公开的又一种实现方式中，该步骤S204可以包括：In yet another implementation of the present disclosure, the step S204 may include:
分别选取对齐后的第一人脸图像和对齐后的第二人脸图像中，与平均形状人脸模型中的设定肤色区域对应的区域为第一肤色区域；In the first people's face image after the alignment and the second people's face image after the alignment are selected respectively, the area corresponding to the set skin color area in the average shape face model is the first skin color area;
分别提取第一肤色区域的肤色特征，并将对齐后的第一人脸图像和对齐后的第二人脸图像中，肤色特征与第一肤色区域的肤色特征相同的区域确定为第二肤色区域；Extract the skin color features of the first skin color area respectively, and determine the area with the same skin color feature as the skin color feature of the first skin color area in the aligned first face image and the aligned second face image as the second skin color area ;
分别将对齐后的第一人脸图像和对齐后的第二人脸图像中，除第一肤色区域和第二肤色区域以外的所有区域，作为对齐后的第一人脸图像和对齐后的第二人脸图像的非肤色区域。In the aligned first human face image and the aligned second human face image, all areas except the first skin color area and the second skin color area are used as the aligned first human face image and the aligned first human face image. The non-skinned area of the two-face image.
可以理解地，人脸中的脸颊部位为肤色区域的概率较大，因此可以将人脸中的脸颊部位作为设定肤色区域。在实际应用中，可以将平均形状人脸模型中脸颊部分的特征点标定为设定肤色区域。It can be understood that the cheek part of the human face is more likely to be a skin color area, so the cheek part of the human face can be used as the set skin color area. In practical applications, the feature points of the cheek part in the average shape face model can be marked as the set skin color area.
在步骤S205中，根据对齐后的第一人脸图像和对齐后的第二人脸图像的非肤色区域，计算第一人脸图像和第二人脸图像的干扰特征对相似度的干扰值。In step S205 , according to the non-skinned area of the aligned first human face image and the aligned second human face image, the interference value of the similarity between the interference features of the first human face image and the second human face image is calculated.
在本实施例中，干扰特征为第一人脸图像和第二人脸图像的人脸区域中特征值相同的非人脸部分的特征。干扰特征可以包括深框眼镜、头发、胡须等，本公开对此不作限制。干扰值用于衡量干扰特征对确定的相似度的干扰程度。In this embodiment, the interference features are features of non-face parts with the same feature value in the face regions of the first face image and the second face image. Interfering features may include deep-frame glasses, hair, beard, etc., without limitation of the present disclosure. The noise value is used to measure how much the noise features interfere with the determined similarity.
在本公开的又一种实现方式中，该步骤S205可以包括：In yet another implementation of the present disclosure, the step S205 may include:
分别将对齐后的第一人脸图像和对齐后的第二人脸图像的非肤色区域的像素的特征值取为1，对齐后的第一人脸图像和对齐后的第二人脸图像的肤色区域的特征值取为0；The eigenvalues of the pixels of the non-skin color area of the first face image after alignment and the second face image after alignment are taken as 1 respectively, and the first face image after alignment and the second face image after alignment The eigenvalue of the skin color area is taken as 0;
对对齐后的第一人脸图像和对齐后的第二人脸图像进行图像交；performing image intersection on the aligned first face image and the aligned second face image;
统计对齐后的第一人脸图像和对齐后的第二人脸图像对应的像素的特征值均为1的像素的数量；Count the number of pixels whose eigenvalues of the pixels corresponding to the first face image after the alignment and the second face image after the alignment are 1;
计算统计得到的数量与对齐后的第一人脸图像或对齐后的第二人脸图像的像素的总数的比值，得到第一人脸图像和第二人脸图像的干扰特征对相似度的干扰值。Calculate the ratio of the number obtained by the statistics to the total number of pixels of the aligned first face image or the aligned second face image to obtain the interference of the interference features of the first face image and the second face image on the similarity value.
在本实施例中，对齐后的第一人脸图像和对齐后的第二人脸图像的肤色区域包括第一肤色区域和第二肤色区域。In this embodiment, the skin color areas of the aligned first human face image and the aligned second human face image include a first skin color area and a second skin color area.
可以理解地，通过依次执行步骤S203-S205，即可实现确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。It can be understood that by performing steps S203-S205 in sequence, the determination of the interference value of the similarity between the interference features of the first face image and the specified face image can be realized.
在步骤S206中，根据干扰值，调整相似度。In step S206, the similarity is adjusted according to the interference value.
在本公开的又一种实现方式中，该步骤S206可以包括：In yet another implementation of the present disclosure, the step S206 may include:
按照预定的函数关系，根据干扰值，确定相似度的修正值；Determine the correction value of the similarity according to the predetermined functional relationship and according to the interference value;
将相似度减去修正值，得到调整后的相似度。The correction value is subtracted from the similarity to obtain the adjusted similarity.
例如，统计得到的数量与第一人脸图像或第二人脸图像的像素的总数的比值为80％，相似度为90分，则将修正值定为50分，第一人脸图像和第二人脸图像的调整后的相似度为90分减去50分，即40分。又如，统计得到的数量与第一人脸图像或第二人脸图像的像素的总数的比值为20％，相似度为80分，则将修正值定为10分，第一人脸图像和第二人脸图像的调整后的相似度为80分减去10分，即70分。For example, the ratio of the quantity obtained by statistics to the total number of pixels of the first face image or the second face image is 80%, and the similarity is 90 points, then the correction value is set as 50 points, the first face image and the second face image The adjusted similarity of two face images is 90 points minus 50 points, that is, 40 points. For another example, the ratio of the quantity obtained by statistics to the total number of pixels of the first face image or the second face image is 20%, and the similarity is 80 points, then the correction value is set as 10 points, the first face image and The adjusted similarity of the second face image is 80 points minus 10 points, that is, 70 points.
本公开实施例通过确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值，并根据该干扰值调整第一人脸图像和指定人脸图像的相似度，避免由于第一人脸图像和指定人脸图像中人脸上均有深框眼镜，或者第一人脸图像和指定人脸图像中人脸具有相同或相似的发型等原因，造成的将相似度不是很高的两个人脸误判为相似度较高，提高了识别的准确率。In the embodiment of the present disclosure, by determining the interference value of the similarity between the interference features of the first face image and the designated face image, and adjusting the similarity between the first face image and the designated face image according to the interference value, avoiding the interference caused by the first face image The similarity is not very high due to reasons such as the face image and the specified face image have deep-frame glasses on the face, or the faces in the first face image and the specified face image have the same or similar hairstyles, etc. The two faces are misjudged as having a higher similarity, which improves the accuracy of recognition.
图3是根据一示例性实施例示出的一种人脸识别装置的框图，适用于判断两个人脸图像的相似度，如图3所示，该装置包括获取模块301、识别模块302、干扰确定模块303和修正模块304。Fig. 3 is a block diagram of a face recognition device shown according to an exemplary embodiment, which is suitable for judging the similarity of two face images. As shown in Fig. 3, the device includes an acquisition module 301, a recognition module 302, an interference determination Module 303 and Modification Module 304 .
该获取模块301被配置为获取第一人脸图像。The acquiring module 301 is configured to acquire a first face image.
该识别模块302被配置为确定第一人脸图像和指定人脸图像的相似度。The recognition module 302 is configured to determine the similarity between the first face image and the designated face image.
该干扰确定模块303被配置为确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。The interference determination module 303 is configured to determine an interference value of the similarity between the interference features of the first face image and the designated face image.
该修正模块304被配置为根据干扰值，调整相似度。The correction module 304 is configured to adjust the similarity according to the interference value.
本公开实施例通过确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值，并根据该干扰值调整第一人脸图像和指定人脸图像的相似度，避免由于第一人脸图像和指定人脸图像中人脸上均有深框眼镜，或者第一人脸图像和指定人脸图像中人脸具有相同或相似的发型等原因，造成的将相似度不是很高的两个人脸误判为相似度较高，提高了识别的准确率。In the embodiment of the present disclosure, by determining the interference value of the similarity between the interference features of the first face image and the designated face image, and adjusting the similarity between the first face image and the designated face image according to the interference value, avoiding the interference caused by the first face image The similarity is not very high due to reasons such as the face image and the specified face image have deep-frame glasses on the face, or the faces in the first face image and the specified face image have the same or similar hairstyles, etc. The two faces are misjudged as having a higher similarity, which improves the accuracy of recognition.
图4是根据一示例性实施例示出的另一种人脸识别装置的框图，适用于判断两个人脸图像的相似度，如图4所示，该装置包括获取模块401、识别模块402、干扰确定模块403和修正模块404。Fig. 4 is a block diagram of another face recognition device shown according to an exemplary embodiment, which is suitable for judging the similarity of two face images. As shown in Fig. 4, the device includes an acquisition module 401, a recognition module 402, an interference A determination module 403 and a correction module 404 .
该获取模块401被配置为获取第一人脸图像。The obtaining module 401 is configured to obtain a first face image.
该识别模块402被配置为确定第一人脸图像和指定人脸图像的相似度。The recognition module 402 is configured to determine the similarity between the first face image and the designated face image.
该干扰确定模块403被配置为确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。The interference determination module 403 is configured to determine an interference value of the similarity between the interference features of the first face image and the designated face image.
该修正模块404被配置为根据干扰值，调整相似度。The correction module 404 is configured to adjust the similarity according to the interference value.
在本实施例的一种实现方式中，该干扰确定模块403可以包括对齐单元4031、区域确定单元4032和干扰计算单元4033。In an implementation manner of this embodiment, the interference determination module 403 may include an alignment unit 4031 , an area determination unit 4032 and an interference calculation unit 4033 .
该对齐单元4031被配置为将第一人脸图像和指定人脸图像分别与设定的平均形状模型对齐。The alignment unit 4031 is configured to align the first face image and the designated face image with the set average shape model respectively.
该区域确定单元4032被配置为分别对对齐后的第一人脸图像和对齐后的指定人脸图像进行肤色分析，确定对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域。The area determining unit 4032 is configured to perform skin color analysis on the aligned first face image and the aligned designated face image respectively, and determine the non-skin color of the aligned first human face image and the aligned designated face image area.
该干扰计算单元4033被配置为根据对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域，计算第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。The interference calculation unit 4033 is configured to calculate the interference value of the interference feature of the first human face image and the specified human face image to the similarity according to the aligned first human face image and the non-skin color area of the aligned specified human face image .
该区域确定单元4032可以用于，The area determining unit 4032 can be used for,
选取对齐后的第一人脸图像和对齐后的指定人脸图像中，与平均形状人脸模型中的设定肤色区域对应的区域为第一肤色区域；In the first people's face image after selecting alignment and the specified people's face image after alignment, the area corresponding to the set skin color area in the average shape face model is the first skin color area;
提取第一肤色区域的肤色特征，并将对齐后的第一人脸图像和对齐后的指定人脸图像中，肤色特征与第一肤色区域的肤色特征相同的区域确定为第二肤色区域；Extracting the skin color feature of the first skin color area, and determining the area with the same skin color feature as the skin color feature of the first skin color area in the aligned first human face image and the aligned designated face image as the second skin color area;
将对齐后的第一人脸图像和对齐后的指定人脸图像中，除第一肤色区域和第二肤色区域以外的所有区域，作为对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域。In the aligned first human face image and the aligned specified human face image, all areas except the first skin color area and the second skin color area are used as the aligned first human face image and the aligned specified human face Non-skinned areas of the image.
该干扰计算单元4033可以用于，The interference calculation unit 4033 can be used to:
分别将对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域的像素的特征值取为1，对齐后的第一人脸图像和对齐后的指定人脸图像的肤色区域的特征值取为0；The eigenvalues of the pixels of the non-skin color area of the first face image after alignment and the designated face image after alignment are taken as 1, and the skin color area of the first face image after alignment and the designated face image after alignment The eigenvalue of is taken as 0;
对对齐后的第一人脸图像和对齐后的指定人脸图像进行图像交；performing image intersection on the aligned first face image and the aligned specified face image;
统计对齐后的第一人脸图像和对齐后的指定人脸图像对应的像素的特征值均为1的像素的数量；Count the number of pixels whose eigenvalues of the first face image after the alignment and the pixels corresponding to the designated face image after the alignment are 1;
计算统计得到的数量与对齐后的第一人脸图像或对齐后的指定人脸图像的像素的总数的比值，得到第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。Calculate the ratio of the number obtained by statistics to the total number of pixels of the aligned first face image or the aligned designated face image to obtain the interference value of the similarity between the interference features of the first face image and the designated face image.
在本实施例的另一种实现方式中，该修正模块404可以包括修正值确定单元4041和分数计算单元4042。In another implementation manner of this embodiment, the correction module 404 may include a correction value determination unit 4041 and a score calculation unit 4042 .
该修正值确定单元4041被配置为按照预定的函数关系，根据干扰值，确定相似度的修正值。The correction value determination unit 4041 is configured to determine a correction value of the similarity according to the interference value according to a predetermined functional relationship.
该分数计算单元4042被配置为将相似度减去修正值，得到调整后的相似度。The score calculation unit 4042 is configured to subtract the correction value from the similarity to obtain the adjusted similarity.
本公开实施例通过确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值，并根据该干扰值调整第一人脸图像和指定人脸图像的相似度，避免由于第一人脸图像和指定人脸图像中人脸上均有深框眼镜，或者第一人脸图像和指定人脸图像中人脸具有相同或相似的发型等原因，造成的将相似度不是很高的两个人脸误判为相似度较高，提高了识别的准确率。In the embodiment of the present disclosure, by determining the interference value of the similarity between the interference features of the first face image and the designated face image, and adjusting the similarity between the first face image and the designated face image according to the interference value, avoiding the interference caused by the first face image The similarity is not very high due to reasons such as the face image and the specified face image have deep-frame glasses on the face, or the faces in the first face image and the specified face image have the same or similar hairstyles, etc. The two faces are misjudged as having a higher similarity, which improves the accuracy of recognition.
关于上述实施例中的装置，其中各个模块执行操作的具体方式已经在有关该方法的实施例中进行了详细描述，此处将不做详细阐述说明。Regarding the apparatus in the foregoing embodiments, the specific manner in which each module executes operations has been described in detail in the embodiments related to the method, and will not be described in detail here.
图5是根据一示例性实施例示出的一种用于人脸识别方法的装置800的框图。例如，装置800可以是移动电话，计算机，数字广播终端，消息收发设备，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等。Fig. 5 is a block diagram of an apparatus 800 for a face recognition method according to an exemplary embodiment. For example, the apparatus 800 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and the like.
参照图5，装置800可以包括以下一个或多个组件：处理组件802，存储器804，电源组件806，多媒体组件808，音频组件810，I/O(Input/Output，输入/输出)的接口812，传感器组件814，以及通信组件816。5, device 800 may include one or more of the following components: processing component 802, memory 804, power supply component 806, multimedia component 808, audio component 810, I/O (Input/Output, input/output) interface 812, sensor component 814, and communication component 816.
处理组件802通常控制装置800的整体操作，诸如与显示，电话呼叫，数据通信，相机操作和记录操作相关联的操作。处理元件802可以包括一个或多个处理器820来执行指令，以完成上述的方法的全部或部分步骤。此外，处理组件802可以包括一个或多个模块，便于处理组件802和其他组件之间的交互。例如，处理部件802可以包括多媒体模块，以方便多媒体组件808和处理组件802之间的交互。The processing component 802 generally controls the overall operations of the device 800, such as those associated with display, telephone calls, data communications, camera operations, and recording operations. The processing element 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above method. Additionally, processing component 802 may include one or more modules that facilitate interaction between processing component 802 and other components. For example, processing component 802 may include a multimedia module to facilitate interaction between multimedia component 808 and processing component 802 .
存储器804被配置为存储各种类型的数据以支持在设备800的操作。这些数据的示例包括用于在装置800上操作的任何应用程序或方法的指令，联系人数据，电话簿数据，消息，图片，视频等。存储器804可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如SRAM(Static Random AccessMemory，静态随机存取存储器)，EEPROM(Electrically Erasable ProgrammableRead-Only Memory，电可擦除可编程只读存储器)，EPROM(ErasableProgrammable Read Only Memory，可擦除可编程只读存储器)，PROM(Programmable Read-Only Memory，可编程只读存储器)，ROM(Read-OnlyMemory，只读存储器)，磁存储器，快闪存储器，磁盘或光盘。The memory 804 is configured to store various types of data to support operations at the device 800 . Examples of such data include instructions for any application or method operating on device 800, contact data, phonebook data, messages, pictures, videos, and the like. Memory 804 can be realized by any type of volatile or non-volatile storage device or their combination, such as SRAM (Static Random Access Memory, static random access memory), EEPROM (Electrically Erasable Programmable Read-Only Memory, electrically erasable Programmable Read-Only Memory), EPROM (ErasableProgrammable Read Only Memory, Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory, Programmable Read-Only Memory), ROM (Read-OnlyMemory, Read-Only Memory), Magnetic memory, flash memory, magnetic disk or optical disk.
电力组件806为装置800的各种组件提供电力。电力组件806可以包括电源管理系统，一个或多个电源，及其他与为装置800生成、管理和分配电力相关联的组件。Power component 806 provides power to various components of device 800 . Power components 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for device 800 .
多媒体组件808包括在该装置800和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括LCD(Liquid Crystal Display，液晶显示器)和TP(Touch Panel，触摸面板)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。该触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与该触摸或滑动操作相关的持续时间和压力。在一些实施例中，多媒体组件808包括一个前置摄像头和/或后置摄像头。当设备800处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。The multimedia component 808 includes a screen that provides an output interface between the device 800 and the user. In some embodiments, the screen may include LCD (Liquid Crystal Display, liquid crystal display) and TP (Touch Panel, touch panel). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or swipe action, but also detect the duration and pressure associated with the touch or swipe action. In some embodiments, the multimedia component 808 includes a front camera and/or a rear camera. When the device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have focal length and optical zoom capability.
音频组件810被配置为输出和/或输入音频信号。例如，音频组件810包括一个MIC(Microphone，麦克风)，当装置800处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风被配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器804或经由通信组件816发送。在一些实施例中，音频组件810还包括一个扬声器，用于输出音频信号。The audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a MIC (Microphone, microphone). When the device 800 is in operation modes, such as calling mode, recording mode and voice recognition mode, the microphone is configured to receive external audio signals. Received audio signals may be further stored in memory 804 or sent via communication component 816 . In some embodiments, the audio component 810 also includes a speaker for outputting audio signals.
I/O接口812为处理组件802和外围接口模块之间提供接口，上述外围接口模块可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。The I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module, which may be a keyboard, a click wheel, a button, and the like. These buttons may include, but are not limited to: a home button, volume buttons, start button, and lock button.
传感器组件814包括一个或多个传感器，用于为装置800提供各个方面的状态评估。例如，传感器组件814可以检测到设备800的打开/关闭状态，组件的相对定位，例如该组件为装置800的显示器和小键盘，传感器组件814还可以检测装置800或装置800一个组件的位置改变，用户与装置800接触的存在或不存在，装置800方位或加速/减速和装置800的温度变化。传感器组件814可以包括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件814还可以包括光传感器，如CMOS(Complementary Metal OxideSemiconductor，互补金属氧化物半导体)或CCD(Charge-coupled Device，电荷耦合元件)图像传感器，用于在成像应用中使用。在一些实施例中，该传感器组件814还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。Sensor assembly 814 includes one or more sensors for providing status assessments of various aspects of device 800 . For example, the sensor component 814 can detect the open/closed state of the device 800, the relative positioning of components, such as the display and keypad of the device 800, the sensor component 814 can also detect a change in the position of the device 800 or a component of the device 800, Presence or absence of user contact with device 800 , device 800 orientation or acceleration/deceleration and temperature changes of device 800 . Sensor assembly 814 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact. The sensor component 814 may also include an optical sensor, such as a CMOS (Complementary Metal Oxide Semiconductor, Complementary Metal Oxide Semiconductor) or a CCD (Charge-coupled Device, Charge-Coupled Device) image sensor, for use in imaging applications. In some embodiments, the sensor component 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor.
通信组件816被配置为便于装置800和其他设备之间有线或无线方式的通信。装置800可以接入基于通信标准的无线网络，如WiFi(Wireless Fidelity，无线保真技术)，2G(Second Generation mobile communication technology，第二代移动通讯技术)或3G(3rd Generation mobile communication technology，第三代移动通讯技术)，或它们的组合。在一个示例性实施例中，通信部件816经由广播信道接收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中，该通信部件816还包括NFC(Near Field Communication,近场通信)模块，以促进短程通信。例如，在NFC模块可基于RFID(Radio FrequencyIdentification，射频识别)技术，IrDA(Infrared Data Association，红外数据协会)技术，UWB(Ultra Wideband，超宽带)技术，BT(Blue Tooth，蓝牙)技术和其他技术来实现。The communication component 816 is configured to facilitate wired or wireless communication between the apparatus 800 and other devices. The device 800 can access wireless networks based on communication standards, such as WiFi (Wireless Fidelity, wireless fidelity technology), 2G (Second Generation mobile communication technology, second generation mobile communication technology) or 3G (3rd Generation mobile communication technology, third Generation mobile communication technology), or their combination. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 further includes an NFC (Near Field Communication, near field communication) module to facilitate short-range communication. For example, the NFC module can be based on RFID (Radio Frequency Identification, radio frequency identification) technology, IrDA (Infrared Data Association, infrared data association) technology, UWB (Ultra Wideband, ultra-wideband) technology, BT (Blue Tooth, Bluetooth) technology and other technologies to fulfill.
在示例性实施例中，装置800可以被一个或多个ASIC(Application SpecificIntegrated Circuit，应用专用集成电路)、DSP(Digital Signal Processing，数字信号处理器)、DSPD(Digital Signal Processing Device，数字信号处理设备)、PLD(Programmable Logic Device，可编程逻辑器件)、FPGA(Field－ProgrammableGate Array，现场可编程门阵列)、控制器、微控制器、微处理器或其他电子元件实现，用于执行上述方法。In an exemplary embodiment, the device 800 may be implemented by one or more ASIC (Application Specific Integrated Circuit, application-specific integrated circuit), DSP (Digital Signal Processing, digital signal processor), DSPD (Digital Signal Processing Device, digital signal processing device ), PLD (Programmable Logic Device, programmable logic device), FPGA (Field-Programmable Gate Array, field programmable gate array), controller, microcontroller, microprocessor or other electronic components to implement the above method.
在示例性实施例中，还提供了一种包括指令的非临时性计算机可读存储介质，例如包括指令的存储器804，上述指令可由装置800的处理器820执行以完成上述方法。例如，该非临时性计算机可读存储介质可以是ROM、RAM(RamdomAccess Memory，随机存取存储器)、CD-ROM(Compact Disc Read-Only Memory，光盘只读存储器)、磁带、软盘和光数据存储设备等。In an exemplary embodiment, there is also provided a non-transitory computer-readable storage medium including instructions, such as the memory 804 including instructions, which can be executed by the processor 820 of the device 800 to implement the above method. For example, the non-transitory computer-readable storage medium can be ROM, RAM (Ramdom Access Memory, random access memory), CD-ROM (Compact Disc Read-Only Memory, optical disk read-only memory), magnetic tape, floppy disk and optical data storage device wait.
一种非临时性计算机可读存储介质，当该存储介质中的指令由终端(智能电视)的处理器执行时，使得终端能执行一种人脸识别方法，该方法包括：A non-transitory computer-readable storage medium, when the instructions in the storage medium are executed by the processor of the terminal (smart TV), the terminal can execute a face recognition method, the method comprising:
获取第一人脸图像；Obtain the first face image;
确定第一人脸图像和指定人脸图像的相似度；Determine the similarity between the first face image and the designated face image;
确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值；Determine the interference value of the first face image and the interference feature of the designated face image to the similarity;
根据干扰值，调整相似度。According to the noise value, the similarity is adjusted.
在本实施例的一种实现方式中，确定第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值，包括：In an implementation of this embodiment, determining the interference value of the interference features of the first face image and the specified face image to the similarity includes:
将第一人脸图像和指定人脸图像分别与设定的平均形状模型对齐；Align the first face image and the specified face image with the set average shape model respectively;
分别对对齐后的第一人脸图像和对齐后的指定人脸图像进行肤色分析，确定对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域；Carry out skin color analysis to the first aligned human face image and the aligned designated human face image respectively, and determine the non-skin color area of the aligned first human face image and the aligned designated human face image;
根据对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域，计算第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。According to the aligned first human face image and the non-skin color area of the aligned specified human face image, calculate the interference value of the similarity between the interference features of the first human face image and the specified human face image.
在本实施例的另一种实现方式中，分别对对齐后的第一人脸图像和对齐后的指定人脸图像进行肤色分析，确定对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域，包括：In another implementation of this embodiment, skin color analysis is performed on the aligned first human face image and the aligned specified human face image respectively, and the aligned first human face image and the aligned specified human face are determined. Non-skinned areas of the image, including:
选取对齐后的第一人脸图像和对齐后的指定人脸图像中，与平均形状人脸模型中的设定肤色区域对应的区域为第一肤色区域；In the first people's face image after selecting alignment and the specified people's face image after alignment, the area corresponding to the set skin color area in the average shape face model is the first skin color area;
提取第一肤色区域的肤色特征，并将对齐后的第一人脸图像和对齐后的指定人脸图像中，肤色特征与第一肤色区域的肤色特征相同的区域确定为第二肤色区域；Extracting the skin color feature of the first skin color area, and determining the area with the same skin color feature as the skin color feature of the first skin color area in the aligned first human face image and the aligned designated face image as the second skin color area;
将对齐后的第一人脸图像和对齐后的指定人脸图像中，除第一肤色区域和第二肤色区域以外的所有区域，作为对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域。In the aligned first human face image and the aligned specified human face image, all areas except the first skin color area and the second skin color area are used as the aligned first human face image and the aligned specified human face Non-skinned areas of the image.
在本实施例的又一种实现方式中，根据对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域，计算第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值，包括：In yet another implementation of this embodiment, according to the aligned first human face image and the non-skin color region of the aligned specified human face image, the interference feature pair of the first human face image and the specified human face image is calculated to be similar degrees of disturbance, including:
分别将对齐后的第一人脸图像和对齐后的指定人脸图像的非肤色区域的像素的特征值取为1，对齐后的第一人脸图像和对齐后的指定人脸图像的肤色区域的特征值取为0；The eigenvalues of the pixels of the non-skin color area of the first face image after alignment and the designated face image after alignment are taken as 1, and the skin color area of the first face image after alignment and the designated face image after alignment The eigenvalue of is taken as 0;
对对齐后的第一人脸图像和对齐后的指定人脸图像进行图像交；performing image intersection on the aligned first face image and the aligned specified face image;
统计对齐后的第一人脸图像和对齐后的指定人脸图像对应的像素的特征值均为1的像素的数量；Count the number of pixels whose eigenvalues of the first face image after the alignment and the pixels corresponding to the designated face image after the alignment are 1;
计算统计得到的数量与对齐后的第一人脸图像或对齐后的指定人脸图像的像素的总数的比值，得到第一人脸图像和指定人脸图像的干扰特征对相似度的干扰值。Calculate the ratio of the number obtained by statistics to the total number of pixels of the aligned first face image or the aligned designated face image to obtain the interference value of the similarity between the interference features of the first face image and the designated face image.
在本实施例的又一种实现方式中，根据干扰值，调整相似度，包括：In yet another implementation of this embodiment, adjusting the similarity according to the interference value includes:
按照预定的函数关系，根据干扰值，确定相似度的修正值；Determine the correction value of the similarity according to the predetermined functional relationship and according to the interference value;
将相似度减去修正值，得到调整后的相似度。The correction value is subtracted from the similarity to obtain the adjusted similarity.
本领域技术人员在考虑说明书及实践这里公开的发明后，将容易想到本发明的其它实施方案。本申请旨在涵盖本发明的任何变型、用途或者适应性变化，这些变型、用途或者适应性变化遵循本发明的一般性原理并包括本公开未公开的本技术领域中的公知常识或惯用技术手段。说明书和实施例仅被视为示例性的，本发明的真正范围和精神由下面的权利要求指出。Other embodiments of the invention will be readily apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any modification, use or adaptation of the present invention, these modifications, uses or adaptations follow the general principles of the present invention and include common knowledge or conventional technical means in the technical field not disclosed in this disclosure . The specification and examples are to be considered exemplary only, with a true scope and spirit of the invention being indicated by the following claims.
应当理解的是，本发明并不局限于上面已经描述并在附图中示出的精确结构，并且可以在不脱离其范围进行各种修改和改变。本发明的范围仅由所附的权利要求来限制。It should be understood that the present invention is not limited to the precise constructions which have been described above and shown in the accompanying drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410240747.4ACN104077563B (en) | 2014-05-30 | 2014-05-30 | Face identification method and device |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410240747.4ACN104077563B (en) | 2014-05-30 | 2014-05-30 | Face identification method and device |
| Publication Number | Publication Date |
|---|---|
| CN104077563Atrue CN104077563A (en) | 2014-10-01 |
| CN104077563B CN104077563B (en) | 2018-02-02 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410240747.4AExpired - Fee RelatedCN104077563B (en) | 2014-05-30 | 2014-05-30 | Face identification method and device |
| Country | Link |
|---|---|
| CN (1) | CN104077563B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104899579A (en)* | 2015-06-29 | 2015-09-09 | 小米科技有限责任公司 | Face recognition method and face recognition device |
| CN106919884A (en)* | 2015-12-24 | 2017-07-04 | 北京汉王智远科技有限公司 | Human facial expression recognition method and device |
| CN106980819A (en)* | 2017-03-03 | 2017-07-25 | 竹间智能科技（上海）有限公司 | Similarity judgement system based on human face five-sense-organ |
| CN107506687A (en)* | 2017-07-17 | 2017-12-22 | 广东欧珀移动通信有限公司 | Liveness detection method and related products |
| CN107683499A (en)* | 2015-06-03 | 2018-02-09 | 索尼公司 | Based on observation identification user and perform the information processor of the function based on user, methods and procedures |
| CN109272006A (en)* | 2017-07-18 | 2019-01-25 | 北京柯斯元科技有限公司 | Anti-counterfeit sign with random texture pattern decision-making system and determination method |
| WO2020258119A1 (en)* | 2019-06-27 | 2020-12-30 | 深圳市汇顶科技股份有限公司 | Face recognition method and apparatus, and electronic device |
| WO2021012647A1 (en)* | 2019-07-24 | 2021-01-28 | 创新先进技术有限公司 | Face verification method and apparatus, server and readable storage medium |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102034079A (en)* | 2009-09-24 | 2011-04-27 | 汉王科技股份有限公司 | Method and system for identifying faces shaded by eyeglasses |
| CN103020579A (en)* | 2011-09-22 | 2013-04-03 | 上海银晨智能识别科技有限公司 | Face recognition method and system, and removing method and device for glasses frame in face image |
| CN103136533A (en)* | 2011-11-28 | 2013-06-05 | 汉王科技股份有限公司 | Face recognition method and device based on dynamic threshold value |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102034079A (en)* | 2009-09-24 | 2011-04-27 | 汉王科技股份有限公司 | Method and system for identifying faces shaded by eyeglasses |
| CN103020579A (en)* | 2011-09-22 | 2013-04-03 | 上海银晨智能识别科技有限公司 | Face recognition method and system, and removing method and device for glasses frame in face image |
| CN103136533A (en)* | 2011-11-28 | 2013-06-05 | 汉王科技股份有限公司 | Face recognition method and device based on dynamic threshold value |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107683499A (en)* | 2015-06-03 | 2018-02-09 | 索尼公司 | Based on observation identification user and perform the information processor of the function based on user, methods and procedures |
| CN104899579A (en)* | 2015-06-29 | 2015-09-09 | 小米科技有限责任公司 | Face recognition method and face recognition device |
| CN106919884A (en)* | 2015-12-24 | 2017-07-04 | 北京汉王智远科技有限公司 | Human facial expression recognition method and device |
| CN106980819A (en)* | 2017-03-03 | 2017-07-25 | 竹间智能科技（上海）有限公司 | Similarity judgement system based on human face five-sense-organ |
| CN107506687A (en)* | 2017-07-17 | 2017-12-22 | 广东欧珀移动通信有限公司 | Liveness detection method and related products |
| WO2019015415A1 (en)* | 2017-07-17 | 2019-01-24 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method of living body detection and terminal device |
| CN107506687B (en)* | 2017-07-17 | 2020-01-21 | Oppo广东移动通信有限公司 | Living body detection method and related product |
| RU2731370C1 (en)* | 2017-07-17 | 2020-09-02 | Гуандун Оппо Мобайл Телекоммьюникейшнс Корп., Лтд. | Method of living organism recognition and terminal device |
| US11100348B2 (en) | 2017-07-17 | 2021-08-24 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Method of living body detection and terminal device |
| CN109272006A (en)* | 2017-07-18 | 2019-01-25 | 北京柯斯元科技有限公司 | Anti-counterfeit sign with random texture pattern decision-making system and determination method |
| WO2020258119A1 (en)* | 2019-06-27 | 2020-12-30 | 深圳市汇顶科技股份有限公司 | Face recognition method and apparatus, and electronic device |
| WO2021012647A1 (en)* | 2019-07-24 | 2021-01-28 | 创新先进技术有限公司 | Face verification method and apparatus, server and readable storage medium |
| Publication number | Publication date |
|---|---|
| CN104077563B (en) | 2018-02-02 |
| Publication | Publication Date | Title |
|---|---|---|
| CN104077563B (en) | Face identification method and device | |
| CN106384098B (en) | Head pose detection method, device and terminal based on image | |
| US10007841B2 (en) | Human face recognition method, apparatus and terminal | |
| CN106355573B (en) | Target positioning method and device in pictures | |
| CN106295566A (en) | Facial expression recognizing method and device | |
| CN107025419B (en) | Fingerprint template inputting method and device | |
| WO2021056808A1 (en) | Image processing method and apparatus, electronic device, and storage medium | |
| CN106295499B (en) | Age estimation method and device | |
| WO2021031609A1 (en) | Living body detection method and device, electronic apparatus and storage medium | |
| CN106295515B (en) | Method and device for determining a face area in an image | |
| CN104850852B (en) | Feature vector computational methods and device | |
| CN104850828A (en) | Person identification method and person identification device | |
| CN106204435A (en) | Image processing method and device | |
| CN107944367B (en) | Face key point detection method and device | |
| CN106778531A (en) | Face detection method and device | |
| CN107730448B (en) | Beauty method and device based on image processing | |
| CN107832741A (en) | The method, apparatus and computer-readable recording medium of facial modeling | |
| CN106295511A (en) | Face tracking method and device | |
| CN107692997A (en) | Heart rate detection method and device | |
| CN107463903B (en) | Face key point positioning method and device | |
| CN107886070A (en) | Verification method, device and the equipment of facial image | |
| CN107729880A (en) | Method for detecting human face and device | |
| JP2017521742A (en) | Method and apparatus for acquiring iris image, and iris identification device | |
| CN106557755A (en) | Fingerprint template acquisition methods and device | |
| CN106339695A (en) | Face similarity detection method, device and terminal |
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee | Granted publication date:20180202 |