技术领域technical field
本发明涉及一种特征图像的特征量提取方法，尤其是涉及一种脸部图像的亮度特征量提取方法。The invention relates to a method for extracting feature quantities of feature images, in particular to a method for extracting brightness feature quantities of facial images.
背景技术Background technique
近年来，提出了一种利用摄像头摄取的特征图案，例如脸部，识别人的身份的实用方法。常用的一种匹配方法是，通过例如在预先设定的注册图像和处理目标图像之间进行标准化相关性等识别处理，来计算相似度，并最终根据相似度来判断待识别的人是否是已经预先注册在识别系统中的人。In recent years, a practical method for identifying a person's identity using characteristic patterns captured by a camera, such as a face, has been proposed. A commonly used matching method is to calculate the similarity by, for example, performing recognition processing such as standardized correlation between the pre-set registration image and the processing target image, and finally judge whether the person to be recognized is already A person pre-registered in the identification system.
当在实时的环境下(例如安全系统)提供使用这种匹配方法的人脸识别时，存在以下可能：由于天气或一天中时辰的变化，图像的反差变化会很大，甚至会在图像中产生局部阴影。在实时环境中，图像中的人的外貌会根据光线的变化显著变化，这将极大地影响识别精度。When providing face recognition using this matching method in a real-time environment (such as a security system), there is a possibility that the contrast of the image will vary greatly due to changes in the weather or the time of day, and even produce Partial shadows. In a real-time environment, the appearance of the person in the image will change significantly according to the change of light, which will greatly affect the recognition accuracy.
为了解决这个问题，已经有人提出了将相对于光线的变化更加稳定的特征量从亮度值中提取出来，而不是在图像的亮度值本身上执行识别处理。例如，将每一个比对部位中的每一个像素的亮度值通过与该像素的环绕像素之间的差值进行修正，然后将修正的亮度值定义为该像素的最终亮度值。In order to solve this problem, it has been proposed to extract a feature quantity that is more stable with respect to changes in light from the luminance value, instead of performing recognition processing on the luminance value itself of the image. For example, the luminance value of each pixel in each comparison site is corrected by the difference with the surrounding pixels of the pixel, and then the corrected luminance value is defined as the final luminance value of the pixel.
然而现有技术存在修正方式粗略，最终亮度值不够精确，以及针对每一个像素进行修正，效率太低的缺点。However, the prior art has disadvantages such as rough correction method, inaccurate final luminance value, and low efficiency of correction for each pixel.
发明内容Contents of the invention
为解决上述技术问题，根据本发明的一个方案，本发明提供一种脸部图像的亮度特征量提取方法，其利用脸部图像识别系统进行亮度特征量提取，所述脸部图像识别系统包括：图像存储单元、调整剪切单元、标示单元、亮度特征量提取单元、存储单元，In order to solve the above-mentioned technical problems, according to a solution of the present invention, the present invention provides a method for extracting brightness features of facial images, which uses a facial image recognition system to extract brightness features, and the facial image recognition system includes: Image storage unit, adjustment and cutting unit, marking unit, brightness feature extraction unit, storage unit,
所述脸部图像的亮度特征量提取方法包括以下步骤：The brightness feature quantity extraction method of described facial image comprises the following steps:
输入步骤，将待登记的脸部图像或待识别的脸部图像通过摄像头输入到所述脸部图像识别系统的所述图像存储单元中；input step, inputting the facial image to be registered or the facial image to be recognized into the image storage unit of the facial image recognition system through the camera;
调整步骤，将所存储的脸部图像通过所述调整剪切单元调整并剪切为方向、尺寸符合一定标准的标准脸部图像；An adjustment step, adjusting and cutting the stored facial image into a standard facial image whose direction and size meet certain standards through the adjustment and cutting unit;
部位标示步骤，将所述标准脸部图像通过所述标示单元按照部位进行分割标示；The part marking step is to segment and mark the standard facial image according to the parts through the marking unit;
提取步骤，对所标示的每一部位图像用所述亮度特征量提取单元提取各部位图像的亮度特征量；The extraction step is to use the brightness feature extraction unit to extract the brightness feature of each part of the image for each part of the marked image;
存储步骤，将所述亮度特征量存储在所述存储单元中；a storing step of storing the luminance feature quantity in the storage unit;
其中，所述提取步骤包括：Wherein, the extraction step comprises:
识别步骤，识别各部位图像中的每一个像素点的亮度值；The identification step identifies the brightness value of each pixel in the image of each part;
第一比较步骤，提取部位图像中的一个像素作为目标像素，将与该目标像素相邻的所有像素所作为目标像素环绕像素，计算各目标像素环绕像素与目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的±20％，则将该环绕像素与所述目标像素归为一目标像素组；The first comparison step is to extract a pixel in the part image as the target pixel, and use all the pixels adjacent to the target pixel as the surrounding pixels of the target pixel, and calculate the brightness difference between the surrounding pixels of each target pixel and the target pixel, if If the luminance difference does not exceed ±20% of the luminance of the target pixel, then the surrounding pixels and the target pixel are classified into a target pixel group;
第二比较步骤，将与所述第一比较步骤中得到的所述目标像素组相邻的所有像素作为像素组环绕像素，计算各像素组环绕像素与在所述第一比较步骤中提取的所述目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+1；如果所述亮度差值超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+2；如果所述亮度差值不超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为0；如果所述亮度差值超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为-1；The second comparison step is to use all the pixels adjacent to the target pixel group obtained in the first comparison step as the surrounding pixels of the pixel group, and calculate the difference between the surrounding pixels of each pixel group and the pixels extracted in the first comparison step. The brightness difference between the target pixels, if the brightness difference does not exceed +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +1; if the brightness difference exceeds the +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +2; if the brightness difference does not exceed -50% of the brightness of the target pixel, count the surrounding pixels of the pixel group as 0 ; if the brightness difference exceeds -50% of the brightness of the target pixel, counting the surrounding pixels of the pixel group as -1;
计算步骤，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，由此得到各目标像素的亮度特征量；The calculation step is to accumulate the counts of the surrounding pixels of each pixel group to obtain a final brightness feature value, and distribute the final brightness feature value to each target pixel in the target pixel group, thereby obtaining the brightness feature value of each target pixel ;
确定步骤，顺序提取目标像素组之外的另一个像素，作为目标像素，重复第一比较步骤、第二比较步骤和计算步骤，直到部位图像中的所有像素均获得亮度特征值。The determining step is sequentially extracting another pixel other than the target pixel group as the target pixel, and repeating the first comparison step, the second comparison step and the calculation step until all pixels in the part image obtain brightness feature values.
进一步地，本发明所述的脸部图像的亮度特征量提取方法中，所述图像存储单元包括登记的脸部图像存储子单元和待识别的脸部图像存储子单元。Furthermore, in the method for extracting brightness feature quantities of facial images according to the present invention, the image storage unit includes a registered facial image storage subunit and a facial image storage subunit to be recognized.
进一步地，本发明所述的脸部图像的亮度特征量提取方法中，输入步骤，将待登记的脸部图像或待识别的脸部图像通过摄像头输入到所述脸部图像识别系统的所述图像存储单元中的所述登记的脸部图像存储子单元中；Further, in the method for extracting the brightness feature value of the facial image according to the present invention, the input step is to input the facial image to be registered or the facial image to be recognized into the said facial image recognition system through the camera. In the registered face image storage subunit in the image storage unit;
进一步地，本发明所述的脸部图像的亮度特征量提取方法中，在所述部位标示步骤中，将所述标准脸部图像通过所述标示单元按照部位以数字的方式进行分割标示。Further, in the facial image brightness feature extraction method according to the present invention, in the part labeling step, the standard facial image is digitally segmented and marked by parts by the labeling unit.
进一步地，本发明所述的脸部图像的亮度特征量提取方法中，在所述部位标示步骤，将所述标准脸部图像通过所述标示单元按照部位进行分割标示，所述部位包括脸部的嘴巴、眉毛、眼睛、鼻子、下巴以及颧骨。Further, in the facial image brightness feature extraction method according to the present invention, in the part labeling step, the standard facial image is segmented and marked by the labeling unit according to the parts, and the parts include the face mouth, eyebrows, eyes, nose, chin, and cheekbones.
进一步地，本发明所述的特征图像的亮度特征量提取方法中，在所述计算步骤中，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，并且，如果所述最终亮度特征值不大于+2或不小于-2，则将所述最终亮度特征值设定为0，由此得到各目标像素的亮度特征量。Further, in the method for extracting luminance feature quantities of feature images according to the present invention, in the calculation step, the counts of surrounding pixels of each pixel group are accumulated to obtain a final luminance feature value, and the final luminance feature value is assigned For each target pixel in the target pixel group, and if the final luminance feature value is not greater than +2 or not less than -2, then the final luminance feature value is set to 0, thereby obtaining each target pixel The brightness feature quantity of .
根据本发明的另一个方案，本发明提供一种特征图像的识别方法，包括以下步骤：According to another solution of the present invention, the present invention provides a method for identifying a feature image, comprising the following steps:
登记图像输入步骤，将待登记的特征图像通过摄像头输入到所述特征图像识别系统的所述图像存储单元中；The registration image input step is to input the characteristic image to be registered into the image storage unit of the characteristic image recognition system through the camera;
调整步骤，将所存储的特征图像通过所述调整剪切单元调整并剪切为方向、尺寸符合一定标准的标准特征图像；An adjustment step, adjusting and cutting the stored characteristic image into a standard characteristic image whose direction and size meet certain standards through the adjustment and cutting unit;
部位标示步骤，将所述标准特征图像通过所述标示单元按照部位进行分割标示；The part labeling step is to divide and mark the standard feature image according to the part through the labeling unit;
提取步骤，对所标示的每一部位图像用所述亮度特征量提取单元提取各部位图像的亮度特征量；The extraction step is to use the brightness feature extraction unit to extract the brightness feature of each part of the image for each part of the marked image;
存储步骤，将所述亮度特征量存储在所述存储单元中；a storing step of storing the luminance feature quantity in the storage unit;
其中，所述提取步骤包括：Wherein, the extraction step comprises:
识别步骤，识别各部位图像中的每一个像素点的亮度值；The identification step identifies the brightness value of each pixel in the image of each part;
第一比较步骤，提取部位图像中的一个像素作为目标像素，将与该目标像素相邻的所有像素所作为目标像素环绕像素，计算各目标像素环绕像素与目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的±20％，则将该环绕像素与所述目标像素归为一目标像素组；The first comparison step is to extract a pixel in the part image as the target pixel, and use all the pixels adjacent to the target pixel as the surrounding pixels of the target pixel, and calculate the brightness difference between the surrounding pixels of each target pixel and the target pixel, if If the luminance difference does not exceed ±20% of the luminance of the target pixel, then the surrounding pixels and the target pixel are classified into a target pixel group;
第二比较步骤，将与所述第一比较步骤中得到的所述目标像素组相邻的所有像素作为像素组环绕像素，计算各像素组环绕像素与在所述第一比较步骤中提取的所述目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+1；如果所述亮度差值超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+2；如果所述亮度差值不超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为0；如果所述亮度差值超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为-1；The second comparison step is to use all the pixels adjacent to the target pixel group obtained in the first comparison step as the surrounding pixels of the pixel group, and calculate the difference between the surrounding pixels of each pixel group and the pixels extracted in the first comparison step. The brightness difference between the target pixels, if the brightness difference does not exceed +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +1; if the brightness difference exceeds the +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +2; if the brightness difference does not exceed -50% of the brightness of the target pixel, count the surrounding pixels of the pixel group as 0 ; if the brightness difference exceeds -50% of the brightness of the target pixel, counting the surrounding pixels of the pixel group as -1;
计算步骤，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，由此得到各目标像素的亮度特征量；The calculation step is to accumulate the counts of the surrounding pixels of each pixel group to obtain a final brightness feature value, and distribute the final brightness feature value to each target pixel in the target pixel group, thereby obtaining the brightness feature value of each target pixel ;
确定步骤，顺序提取目标像素组之外的另一个像素，作为目标像素，重复第一比较步骤、第二比较步骤和计算步骤，直到部位图像中的所有像素均获得亮度特征值；A determining step, sequentially extracting another pixel other than the target pixel group as the target pixel, repeating the first comparison step, the second comparison step and the calculation step until all pixels in the part image obtain brightness feature values;
待识别图像输入步骤，将待识别的特征图像通过摄像头输入到所述特征图像识别系统的所述图像存储单元中；The step of inputting the image to be recognized is inputting the characteristic image to be recognized into the image storage unit of the characteristic image recognition system through the camera;
重复上述调整步骤、部位标示步骤、提取步骤以及存储步骤，从而获得待识别的特征图像的亮度特征量；repeating the above adjustment steps, part labeling steps, extraction steps and storage steps, so as to obtain the brightness feature quantity of the feature image to be identified;
根据用户选取的三个以上的特征部位或者根据所述图像识别系统设定的多个部位分别从图像存储单元中提取所述部位的待识别的特征图像的亮度特征量和登记的特征图像的亮度特征量；According to the three or more characteristic parts selected by the user or according to the multiple parts set by the image recognition system, the brightness feature value of the characteristic image to be recognized and the brightness of the registered characteristic image of the part are respectively extracted from the image storage unit Feature amount;
将所提取的所述待识别的特征图像的亮度特征量与所述登记的特征图像的亮度特征量进行比较，如果两者类似，则判定为待识别的特征图像通过识别；否则如果两者不类似，则判定为待识别的特征图像未通过识别。Comparing the extracted luminance feature value of the feature image to be recognized with the luminance feature value of the registered feature image, if the two are similar, it is determined that the feature image to be recognized has passed the recognition; otherwise, if the two are not Similarly, it is determined that the feature image to be recognized has not passed the recognition.
进一步地，本发明所述的特征图像的识别方法中，所述用户选取的三个以上的特征部位中至少包括脸部的嘴巴、眉毛、眼睛、鼻子、下巴以及颧骨其中之一。Furthermore, in the feature image recognition method of the present invention, the three or more feature parts selected by the user include at least one of the mouth, eyebrows, eyes, nose, chin, and cheekbones of the face.
进一步地，本发明所述的特征图像的亮度特征量提取方法中，在所述计算步骤中，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，并且，如果所述最终亮度特征值不大于+2或不小于-2，则将所述最终亮度特征值设定为0，由此得到各目标像素的亮度特征量。Further, in the method for extracting luminance feature quantities of feature images according to the present invention, in the calculation step, the counts of surrounding pixels of each pixel group are accumulated to obtain a final luminance feature value, and the final luminance feature value is assigned For each target pixel in the target pixel group, and if the final luminance feature value is not greater than +2 or not less than -2, then the final luminance feature value is set to 0, thereby obtaining each target pixel The brightness feature quantity of .
进一步地，如权利要求7所述的特征图像的亮度特征量提取方法，其中，在所述第一比较步骤中，如果所述亮度差值不超过所述目标像素的亮度的±10％时，则将该环绕像素与所述目标像素归为一目标像素组。Further, the method for extracting the brightness feature value of the characteristic image according to claim 7, wherein, in the first comparison step, if the brightness difference does not exceed ±10% of the brightness of the target pixel, Then the surrounding pixels and the target pixels are classified into a target pixel group.
通过本发明所提供的技术方案，识别系统可以将亮度类似的像素归为一组，从而提高了效率。并且根据环绕像素和目标像素的亮度差值，更加细致地为环绕像素设定参数值，从而更加精确地确定目标像素的最终亮度特征值，从而得到各目标像素的亮度特征量。Through the technical solution provided by the invention, the recognition system can group pixels with similar brightness into one group, thereby improving efficiency. And according to the brightness difference between the surrounding pixel and the target pixel, the parameter value is set for the surrounding pixel more carefully, so as to more accurately determine the final brightness characteristic value of the target pixel, thereby obtaining the brightness characteristic value of each target pixel.
具体实施方式Detailed ways
下面将通过实施例详细说明本发明：The present invention will be described in detail below by embodiment:
本发明提供的特征图像的亮度特征量提取方法中利用了特征图像识别系统。特征识别系统可以通过对已经存储的特征特征和待检验的特征特征进行比对，而确定待识别的人是否是已经注册的人员中的某一个。所述特征图像识别系统包括：图像存储单元、调整剪切单元、标示单元、亮度特征量提取单元、存储单元。这些单元的作用如下所述。The characteristic image recognition system is utilized in the method for extracting the brightness feature quantity of the characteristic image provided by the present invention. The feature recognition system can determine whether the person to be identified is one of the registered persons by comparing the stored feature with the feature to be checked. The feature image recognition system includes: an image storage unit, an adjustment and cutting unit, a labeling unit, a brightness feature extraction unit, and a storage unit. The functions of these units are described below.
然而，通过摄像头取得的特征图像会因肤色以及因外界光线的变化而产生亮度差异，这种差异可能会导致识别的失败。因此需要特别重视特征图像的亮度特征，因此需要提供一种最大可能排除光线变化影响的特征图像的亮度特征量提取方法。However, the feature image obtained by the camera will have brightness differences due to changes in skin color and external light, and this difference may cause recognition failures. Therefore, it is necessary to pay special attention to the brightness feature of the feature image, so it is necessary to provide a brightness feature extraction method of the feature image that can eliminate the influence of light changes to the greatest extent.
本发明提供的特征图像的亮度特征量提取方法包括以下步骤：The brightness feature quantity extraction method of feature image provided by the invention comprises the following steps:
输入步骤，将待登记的特征图像或待识别的特征图像通过摄像头输入到所述特征图像识别系统的所述图像存储单元中。由于这种方法既适于对预先录入的登记的特征图像进行处理，又适合对应用时需要进行识别的特征图像进行处理，因此在所述输入步骤中，输入的可以是两种图像中的任一种。The input step is to input the characteristic image to be registered or the characteristic image to be recognized into the image storage unit of the characteristic image recognition system through the camera. Because this method is not only suitable for processing the pre-registered feature images, but also suitable for processing the feature images that need to be identified during application, so in the input step, the input can be any of the two images. A sort of.
调整步骤，将所存储的特征图像通过所述调整剪切单元调整并剪切为方向、尺寸符合一定标准的标准特征图像。调整剪切单元对输入的处理目标图像执行检测，并且确定处理目标图像中的特征的位置、大小和方向。此外，调整剪切单元基于所确定的特征位置、特征尺寸和特征方向标准化面部尺寸至预定的尺寸；通过剪切特征图像产生剪切的标准图形，这样，特征部位沿预定的方向取向；以及输出所剪切的标准图像。In the adjustment step, the stored characteristic image is adjusted and cut by the adjustment and cutting unit into a standard characteristic image whose orientation and size conform to a certain standard. The trimming unit performs detection on the input processing target image, and determines the position, size, and orientation of features in the processing target image. In addition, the adjustment clipping unit normalizes the face size to a predetermined size based on the determined feature position, feature size, and feature direction; generates a clipped standard figure by clipping the feature image so that the feature portion is oriented in a predetermined direction; and outputs The standard image that is cropped.
部位标示步骤，将所述标准特征图像通过所述标示单元按照部位进行分割标示。标示的时候可以用数字进行标示，也可以用字母进行标示。标示是为了便于将来进行识别的时候，能够准确地在注册的特征图像和待识的别特征图像上选择出相同的部位。In the part labeling step, the standard feature image is segmented and marked according to parts by the labeling unit. When marking, numbers can be used for marking, and letters can also be used for marking. The purpose of marking is to accurately select the same part on the registered feature image and the unique feature image to be recognized for future recognition.
提取步骤，对所标示的每一部位图像用所述亮度特征量提取单元提取各部位图像的亮度特征量；The extraction step is to use the brightness feature extraction unit to extract the brightness feature of each part of the image for each part of the marked image;
其中，所述提取步骤包括：Wherein, the extraction step comprises:
识别步骤，识别各部位图像中的每一个像素点的亮度值。The recognition step is to recognize the brightness value of each pixel in the image of each part.
第一比较步骤，提取部位图像中的一个像素作为目标像素，将与该目标像素相邻的所有像素所作为目标像素环绕像素，计算各目标像素环绕像素与目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的±20％，则将该环绕像素与所述目标像素归为一目标像素组。例如选定的目标像素的亮度值是100，则环绕像素中亮度值介于80～120之间的可以与所述目标像素归为一组，成为一个目标像素组。假定说，目标像素周围有八个环绕像素，而其中两个环绕像素的亮度值介于80～120之间，那么这两个环绕像素和那个目标像素组成包含三个像素的目标像素组。The first comparison step is to extract a pixel in the part image as the target pixel, and use all the pixels adjacent to the target pixel as the surrounding pixels of the target pixel, and calculate the brightness difference between the surrounding pixels of each target pixel and the target pixel, if If the brightness difference does not exceed ±20% of the brightness of the target pixel, then the surrounding pixels and the target pixel are classified into a target pixel group. For example, if the brightness value of the selected target pixel is 100, surrounding pixels with brightness values ranging from 80 to 120 may be grouped with the target pixel to form a target pixel group. Assume that there are eight surrounding pixels around the target pixel, and the luminance values of two of the surrounding pixels are between 80 and 120, then the two surrounding pixels and the target pixel form a target pixel group including three pixels.
第二比较步骤，将与所述第一比较步骤中得到的所述目标像素组相邻的所有像素作为像素组环绕像素，计算各像素组环绕像素与在所述第一比较步骤中提取的所述目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+1；如果所述亮度差值超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+2；如果所述亮度差值不超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为0；如果所述亮度差值超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为-1。例如，将上面包含三个像素的目标像素组相邻的像素，也就是，和所述三个像素中任意一个相邻的像素作为像素组环绕像素。然而对各个环绕像素设定参数，参数设定规则就是如果所述亮度差值不超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+1；如果所述亮度差值超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+2；如果所述亮度差值不超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为0；如果所述亮度差值超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为-1。这个规则能够使得环绕像素的亮度参数更加精确，从而使得后续的对目标像素亮度参数的设定更为准确。The second comparison step is to use all the pixels adjacent to the target pixel group obtained in the first comparison step as the surrounding pixels of the pixel group, and calculate the difference between the surrounding pixels of each pixel group and the pixels extracted in the first comparison step. The brightness difference between the target pixels, if the brightness difference does not exceed +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +1; if the brightness difference exceeds the +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +2; if the brightness difference does not exceed -50% of the brightness of the target pixel, count the surrounding pixels of the pixel group as 0 ; If the luminance difference exceeds -50% of the luminance of the target pixel, count the surrounding pixels of the pixel group as -1. For example, the pixels adjacent to the target pixel group including three pixels above, that is, the pixels adjacent to any one of the three pixels are used as the surrounding pixels of the pixel group. However, parameters are set for each surrounding pixel, and the parameter setting rule is that if the brightness difference does not exceed +100% of the brightness of the target pixel, the surrounding pixels of the pixel group are counted as +1; If the value exceeds +100% of the brightness of the target pixel, surround the pixel group with a pixel count of +2; if the difference in brightness does not exceed -50% of the brightness of the target pixel, surround the pixel group The pixel count is 0; if the brightness difference exceeds -50% of the brightness of the target pixel, the surrounding pixels of the pixel group are counted as -1. This rule can make the brightness parameters of the surrounding pixels more accurate, thereby making the subsequent setting of the brightness parameters of the target pixels more accurate.
计算步骤，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，由此得到各目标像素的亮度特征量。也就是说，一个目标像素组中的各目标像素的亮度特征值(亮度特征量)是一样的。从而简化了计算过程，提供了识别效率。The calculation step is to accumulate the counts of the surrounding pixels of each pixel group to obtain a final brightness feature value, and distribute the final brightness feature value to each target pixel in the target pixel group, thereby obtaining the brightness feature value of each target pixel . That is, the luminance feature values (luminance feature quantities) of the target pixels in one target pixel group are the same. Thus, the calculation process is simplified and the recognition efficiency is improved.
确定步骤，顺序提取目标像素组之外的另一个像素，作为目标像素，重复第一比较步骤、第二比较步骤和计算步骤，直到部位图像中的所有像素均获得亮度特征值。这样这个部位像素就获得了亮度特征量。The determining step is sequentially extracting another pixel other than the target pixel group as the target pixel, and repeating the first comparison step, the second comparison step and the calculation step until all pixels in the part image obtain brightness feature values. In this way, the pixel in this part obtains the luminance feature quantity.
最后是存储步骤，将所述亮度特征量存储在所述存储单元中。Finally, there is a storage step of storing the luminance feature value in the storage unit.
所述的特征图像的亮度特征量提取方法中，所述图像存储单元包括登记的特征图像存储子单元和待识别的特征图像存储子单元。这样可以将待登记的特征图像和待识别的特征图像分开存储。In the method for extracting brightness feature quantities of feature images, the image storage unit includes a registered feature image storage subunit and a feature image storage subunit to be identified. In this way, the characteristic images to be registered and the characteristic images to be recognized can be stored separately.
所述的特征图像的亮度特征量提取方法中，在所述输入步骤中，将待登记的特征图像或待识别的特征图像通过摄像头输入到所述特征图像识别系统的所述图像存储单元中的所述登记的特征图像存储子单元中。In the method for extracting the brightness feature quantity of the feature image, in the input step, the feature image to be registered or the feature image to be recognized is input into the image storage unit of the feature image recognition system through the camera The registered feature image is stored in the subunit.
所述的特征图像的亮度特征量提取方法中，在所述部位标示步骤中，将所述标准特征图像通过所述标示单元按照部位以数字的方式进行分割标示。实际上也可以以字母的方式进行标示，但是以数字的方式标示，比较容易与亮度特征量标示结合起来。In the method for extracting luminance feature quantities of feature images, in the part labeling step, the standard feature image is digitally segmented and marked by parts by the labeling unit. In fact, it is also possible to mark by letters, but it is easier to combine with the mark of the luminance characteristic value by marking in numbers.
所述的特征图像的亮度特征量提取方法中，在所述部位标示步骤中，将所述标准特征图像通过所述标示单元按照部位进行分割标示，所述部位包括脸部的嘴巴、眉毛、眼睛、鼻子、下巴以及颧骨。通常在进行识别的时候，脸部图像比对的部位主要是眼睛、鼻子和嘴巴。而在一些特殊情况，例如，待识别者戴眼镜或口罩的情况下，也可以由待识别者自行选择想要识别的部位。In the method for extracting the brightness feature quantity of the feature image, in the part labeling step, the standard feature image is segmented and marked by the labeling unit according to the parts, and the parts include the mouth, eyebrows, and eyes of the face. , nose, chin, and cheekbones. Usually, when performing recognition, the main parts of facial image comparison are eyes, nose and mouth. In some special cases, for example, when the person to be recognized wears glasses or a mask, the person to be recognized can also choose the part to be recognized by himself.
所述的特征图像的亮度特征量提取方法中，在所述计算步骤中，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，并且，如果所述最终亮度特征值不大于+2或不小于-2，则将所述最终亮度特征值设定为0，由此得到各目标像素的亮度特征量。当最终亮度特征值不大于+2或不小于-2时，也就是说，当最终亮度特征值是0，1或-1时，可以直接将所述最终亮度特征值设定为0，这说明该目标像素(目标像素组)周围的亮度变化不是很大。In the method for extracting the brightness feature value of the feature image, in the calculation step, the counts of the surrounding pixels of each pixel group are accumulated to obtain a final brightness feature value, and the final brightness feature value is assigned to the target pixel Each target pixel in the group, and if the final luminance feature value is not greater than +2 or not less than -2, then the final luminance feature value is set to 0, thereby obtaining the luminance feature value of each target pixel. When the final luminance feature value is not greater than +2 or not less than -2, that is, when the final luminance feature value is 0, 1 or -1, the final luminance feature value can be directly set to 0, which means The brightness variation around this target pixel (target pixel group) is not very large.
本发明还提供了一种特征图像的识别方法，包括以下步骤：The present invention also provides a method for identifying feature images, comprising the following steps:
登记图像输入步骤，将待登记的特征图像通过摄像头输入到所述特征图像识别系统的所述图像存储单元中；The registration image input step is to input the characteristic image to be registered into the image storage unit of the characteristic image recognition system through the camera;
调整步骤，将所存储的特征图像通过所述调整剪切单元调整并剪切为方向、尺寸符合一定标准的标准特征图像；An adjustment step, adjusting and cutting the stored characteristic image into a standard characteristic image whose direction and size meet certain standards through the adjustment and cutting unit;
部位标示步骤，将所述标准特征图像通过所述标示单元按照部位进行分割标示；The part labeling step is to divide and mark the standard feature image according to the part through the labeling unit;
提取步骤，对所标示的每一部位图像用所述亮度特征量提取单元提取各部位图像的亮度特征量；The extraction step is to use the brightness feature extraction unit to extract the brightness feature of each part of the image for each part of the marked image;
存储步骤，将所述亮度特征量存储在所述存储单元中；a storing step of storing the luminance feature quantity in the storage unit;
其中，所述提取步骤包括：Wherein, the extraction step comprises:
识别步骤，识别各部位图像中的每一个像素点的亮度值；The identification step identifies the brightness value of each pixel in the image of each part;
第一比较步骤，提取部位图像中的一个像素作为目标像素，将与该目标像素相邻的所有像素所作为目标像素环绕像素，计算各目标像素环绕像素与目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的±20％，则将该环绕像素与所述目标像素归为一目标像素组；The first comparison step is to extract a pixel in the part image as the target pixel, and use all the pixels adjacent to the target pixel as the surrounding pixels of the target pixel, and calculate the brightness difference between the surrounding pixels of each target pixel and the target pixel, if If the luminance difference does not exceed ±20% of the luminance of the target pixel, then the surrounding pixels and the target pixel are classified into a target pixel group;
第二比较步骤，将与所述第一比较步骤中得到的所述目标像素组相邻的所有像素作为像素组环绕像素，计算各像素组环绕像素与在所述第一比较步骤中提取的所述目标像素之间的亮度差值，如果所述亮度差值不超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+1；如果所述亮度差值超过所述目标像素的亮度的+100％，则将该像素组环绕像素计数为+2；如果所述亮度差值不超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为0；如果所述亮度差值超过所述目标像素的亮度的-50％，则将该像素组环绕像素计数为-1；The second comparison step is to use all the pixels adjacent to the target pixel group obtained in the first comparison step as the surrounding pixels of the pixel group, and calculate the difference between the surrounding pixels of each pixel group and the pixels extracted in the first comparison step. The brightness difference between the target pixels, if the brightness difference does not exceed +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +1; if the brightness difference exceeds the +100% of the brightness of the target pixel, count the surrounding pixels of the pixel group as +2; if the brightness difference does not exceed -50% of the brightness of the target pixel, count the surrounding pixels of the pixel group as 0 ; if the brightness difference exceeds -50% of the brightness of the target pixel, counting the surrounding pixels of the pixel group as -1;
计算步骤，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，由此得到各目标像素的亮度特征量；The calculation step is to accumulate the counts of the surrounding pixels of each pixel group to obtain a final brightness feature value, and distribute the final brightness feature value to each target pixel in the target pixel group, thereby obtaining the brightness feature value of each target pixel ;
确定步骤，顺序提取目标像素组之外的另一个像素，作为目标像素，重复第一比较步骤、第二比较步骤和计算步骤，直到部位图像中的所有像素均获得亮度特征值；A determining step, sequentially extracting another pixel other than the target pixel group as the target pixel, repeating the first comparison step, the second comparison step and the calculation step until all pixels in the part image obtain brightness feature values;
待识别图像输入步骤，将待识别的特征图像通过摄像头输入到所述特征图像识别系统的所述图像存储单元中；The step of inputting the image to be recognized is inputting the characteristic image to be recognized into the image storage unit of the characteristic image recognition system through the camera;
重复上述调整步骤、部位标示步骤、提取步骤以及存储步骤，从而获得待识别的特征图像的亮度特征量；repeating the above adjustment steps, part labeling steps, extraction steps and storage steps, so as to obtain the brightness feature quantity of the feature image to be identified;
根据用户选取的三个以上的特征部位或者根据所述图像识别系统设定的多个部位分别从图像存储单元中提取所述部位的待识别的特征图像的亮度特征量和登记的特征图像的亮度特征量；According to the three or more characteristic parts selected by the user or according to the multiple parts set by the image recognition system, the brightness feature value of the characteristic image to be recognized and the brightness of the registered characteristic image of the part are respectively extracted from the image storage unit Feature amount;
将所提取的所述待识别的特征图像的亮度特征量与所述登记的特征图像的亮度特征量进行比较，如果两者类似，则判定为待识别的特征图像通过识别；否则如果两者不类似，则判定为待识别的特征图像未通过识别。Comparing the extracted luminance feature value of the feature image to be recognized with the luminance feature value of the registered feature image, if the two are similar, it is determined that the feature image to be recognized has passed the recognition; otherwise, if the two are not Similarly, it is determined that the feature image to be recognized has not passed the recognition.
所述的特征图像的识别方法中，所述用户选取的三个以上的特征部位中至少包括脸部的嘴巴、眉毛、眼睛、鼻子、下巴以及颧骨其中之一。In the feature image recognition method, the three or more feature parts selected by the user include at least one of the mouth, eyebrows, eyes, nose, chin, and cheekbones of the face.
所述的特征图像的亮度特征量提取方法中，在所述计算步骤中，将所述各像素组环绕像素的计数累加得到一最终亮度特征值，将该最终亮度特征值分配给所述目标像素组中的各目标像素，并且，如果所述最终亮度特征值不大于+2或不小于-2，则将所述最终亮度特征值设定为0，由此得到各目标像素的亮度特征量。In the method for extracting the brightness feature value of the feature image, in the calculation step, the counts of the surrounding pixels of each pixel group are accumulated to obtain a final brightness feature value, and the final brightness feature value is assigned to the target pixel Each target pixel in the group, and if the final luminance feature value is not greater than +2 or not less than -2, then the final luminance feature value is set to 0, thereby obtaining the luminance feature value of each target pixel.
所述的特征图像的亮度特征量提取方法中，在所述第一比较步骤中，如果所述亮度差值不超过所述目标像素的亮度的±10％时，则将该环绕像素与所述目标像素归为一目标像素组。In the method for extracting the luminance feature quantity of the feature image, in the first comparison step, if the luminance difference does not exceed ±10% of the luminance of the target pixel, the surrounding pixel is compared with the The target pixels are grouped into a target pixel group.
本发明中所揭示的多种功能结构及其实施例，均通过权利要求得到体现和保护，任何根据本发明中所示的附图和实施例所得到的启示，均落入本发明所保护的范围之中。The various functional structures and embodiments disclosed in the present invention are all embodied and protected by the claims, and any enlightenment obtained according to the drawings and embodiments shown in the present invention shall fall into the protection of the present invention. within range.
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010132229CN101814143B (en) | 2010-03-25 | 2010-03-25 | Extraction method for brightness characteristic quantity of feature image and recognition method for feature image |
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010132229CN101814143B (en) | 2010-03-25 | 2010-03-25 | Extraction method for brightness characteristic quantity of feature image and recognition method for feature image |
| Publication Number | Publication Date |
|---|---|
| CN101814143Atrue CN101814143A (en) | 2010-08-25 |
| CN101814143B CN101814143B (en) | 2013-01-23 |
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201010132229ActiveCN101814143B (en) | 2010-03-25 | 2010-03-25 | Extraction method for brightness characteristic quantity of feature image and recognition method for feature image |
| Country | Link |
|---|---|
| CN (1) | CN101814143B (en) |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115100011A (en)* | 2022-06-16 | 2022-09-23 | 慧之安信息技术股份有限公司 | A marine intelligent monitoring system based on edge computing |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080240573A1 (en)* | 2007-03-30 | 2008-10-02 | Aisin Aw Co., Ltd. | Feature information collecting apparatus and feature information collecting method |
| WO2008150002A1 (en)* | 2007-05-31 | 2008-12-11 | Aisin Aw Co., Ltd. | Feature extraction method, and image recognition method and feature database creation method using the same |
| CN101414357A (en)* | 2008-11-12 | 2009-04-22 | 北京中星微电子有限公司 | Method and apparatus for detecting film flicker |
| CN101447023A (en)* | 2008-12-23 | 2009-06-03 | 北京中星微电子有限公司 | Method and system for detecting human head |
| CN101782967A (en)* | 2010-03-19 | 2010-07-21 | 周庆芬 | Method for extracting brightness characteristic quantity of face image and method for identifying face image |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080240573A1 (en)* | 2007-03-30 | 2008-10-02 | Aisin Aw Co., Ltd. | Feature information collecting apparatus and feature information collecting method |
| WO2008150002A1 (en)* | 2007-05-31 | 2008-12-11 | Aisin Aw Co., Ltd. | Feature extraction method, and image recognition method and feature database creation method using the same |
| CN101414357A (en)* | 2008-11-12 | 2009-04-22 | 北京中星微电子有限公司 | Method and apparatus for detecting film flicker |
| CN101447023A (en)* | 2008-12-23 | 2009-06-03 | 北京中星微电子有限公司 | Method and system for detecting human head |
| CN101782967A (en)* | 2010-03-19 | 2010-07-21 | 周庆芬 | Method for extracting brightness characteristic quantity of face image and method for identifying face image |
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115100011A (en)* | 2022-06-16 | 2022-09-23 | 慧之安信息技术股份有限公司 | A marine intelligent monitoring system based on edge computing |
| Publication number | Publication date |
|---|---|
| CN101814143B (en) | 2013-01-23 |
| Publication | Publication Date | Title |
|---|---|---|
| JP4335565B2 (en) | Method and apparatus for detecting and / or tracking one or more color regions in an image or sequence of images | |
| CN102332093B (en) | Identity authentication method and device adopting palmprint and human face fusion recognition | |
| KR20190097640A (en) | Device and method for matching image | |
| CN105654048A (en) | Multi-visual-angle face comparison method | |
| CN112434578A (en) | Mask wearing normative detection method and device, computer equipment and storage medium | |
| CN110363052B (en) | Method and device for determining human face pose in image and computer equipment | |
| CN110796101A (en) | Face recognition method and system of embedded platform | |
| CN109002799A (en) | Face identification method | |
| CN106599776B (en) | A People Counting Method Based on Trajectory Analysis | |
| CN103150720A (en) | Silhouette correction method and system and silhouette extraction method and system | |
| CN110090425A (en) | A kind of clocking method of middle-distance race campaign | |
| CN105335720B (en) | The acquisition method and acquisition system of iris information | |
| CN104794459A (en) | Video personnel identification method | |
| CN114333023A (en) | Face and gait multimodal weighted fusion identity recognition method and system based on angle estimation | |
| CN107346408A (en) | Age recognition methods based on face feature | |
| CN109711231A (en) | A kind of face alignment method based on average face-Pu Shi transformation | |
| CN101814143B (en) | Extraction method for brightness characteristic quantity of feature image and recognition method for feature image | |
| CN108694353A (en) | A kind of multi-modal personal identification method of recognition of face and iris recognition | |
| CN101859380B (en) | Method for extracting luminance feature quantity of facial image and method for recognizing facial image | |
| CN101782967B (en) | Method for extracting brightness characteristic quantity of face image and method for identifying face image | |
| CN109934143A (en) | A kind of method and apparatus of the detection of iris image Sino-U.S. pupil | |
| CN109635682A (en) | A kind of face identification device and method | |
| EP4120184A1 (en) | Information processing device, information processing method, and program | |
| CN113920556A (en) | Face anti-counterfeiting method and device, storage medium and electronic equipment | |
| CN108875646B (en) | Method and system for double comparison and authentication of real face image and identity card registration |
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C56 | Change in the name or address of the patentee | ||
| CP02 | Change in the address of a patent holder | Address after:536000 the Guangxi Zhuang Autonomous Region Industrial Park, Beihai, No. 5, No. 302, deep blue incubation base room 22 Patentee after:Zhou Qingfen Address before:315010 Haishu province Ningbo District Yin Jiang Village, building 507, room 56, Patentee before:Zhou Qingfen | |
| ASS | Succession or assignment of patent right | Owner name:GUANGDONG LV'AN INDUSTRIAL CO., LTD. Free format text:FORMER OWNER: ZHOU QINGFEN Effective date:20150819 | |
| C41 | Transfer of patent application or patent right or utility model | ||
| C53 | Correction of patent of invention or patent application | ||
| CB03 | Change of inventor or designer information | Inventor after:Fu Changshan Inventor after:Pei Jun Inventor before:Zhou Qingfen | |
| COR | Change of bibliographic data | Free format text:CORRECT: INVENTOR; FROM: ZHOU QINGFEN TO: FU CHANGSHAN PEI JUN | |
| TR01 | Transfer of patent right | Effective date of registration:20150819 Address after:510000 Guangdong city of Guangzhou province Tianhe District Tianhe Software Park Building 3 No. 1035 high way Patentee after:GUANGDONG LVAN INDUSTRY AND COMMERCE CO.,LTD. Address before:536000 the Guangxi Zhuang Autonomous Region Industrial Park, Beihai, No. 5, No. 302, deep blue incubation base room 22 Patentee before:Zhou Qingfen | |
| CP02 | Change in the address of a patent holder | Address after:510000 first floor, No.1, Jinying Road, Tianhe District, Guangzhou City, Guangdong Province Patentee after:GUANGDONG LVAN INDUSTRY AND COMMERCE CO.,LTD. Address before:510000 Guangdong city of Guangzhou province Tianhe District Tianhe Software Park Building 3 No. 1035 high way Patentee before:GUANGDONG LVAN INDUSTRY AND COMMERCE CO.,LTD. | |
| CP02 | Change in the address of a patent holder | ||
| CP03 | Change of name, title or address | Address after:510000 first floor, No.1, Jinying Road, Tianhe District, Guangzhou City, Guangdong Province Patentee after:Lu'an Technology Co.,Ltd. Country or region after:China Address before:510000 first floor, No.1, Jinying Road, Tianhe District, Guangzhou City, Guangdong Province Patentee before:GUANGDONG LVAN INDUSTRY AND COMMERCE CO.,LTD. Country or region before:China | |
| CP03 | Change of name, title or address |