摘要：【中文】本发明公开了一种用于K12教育阶段基于人脸识别的视频中检索人员方法，通过视频拍摄设备，获取视频流数据，把视频数据传输到原始数据储存模块中，通过分布式文件系统对原始视频流数据进行切分，把切分处理后原始视频流数据储存到分布式数据库中,对每一片原始视频流数据的帧进行映射到MapReduce模型中，然后分别提取视频物体的脸部特征和动作特征，最终提取出视频中运动的物体；再通过深度学习方法，筛选出运动的物体为人员的视频,然后通过CNN卷积神经网络深度学习方法反复对比，提取人员的外貌特征；将识别到的体貌特征信息保存到云服务数据库中，由相似度高低排序，将要搜索的视频跟需要搜索的人员信息进行一一对应，最终从视频中识别检索到相应的人员。 【EN】The invention discloses a method for retrieving people in videos based on face recognition in a K12 education stage, which comprises the steps of obtaining video stream data through a video shooting device, transmitting the video stream data to an original data storage module, segmenting the original video stream data through a distributed file system, storing the segmented original video stream data in a distributed database, mapping frames of each piece of original video stream data to a MapReduce model, then respectively extracting facial features and action features of a video object, and finally extracting a moving object in the video; then, screening out videos with moving objects as personnel by a deep learning method, and repeatedly comparing the videos by a CNN convolutional neural network deep learning method to extract the appearance characteristics of the personnel; and storing the identified physical feature information into a cloud service database, sequencing according to the similarity, carrying out one-to-one correspondence on the video to be searched and the information of the personnel to be searched, and finally identifying and retrieving the corresponding personnel from the video.

摘要：【中文】本发明公开一种用于K12教育阶段基于视频影像超低码流无损压缩方法，包括以下步骤：捕获原始视频流；对原始视频流编码，运用边缘滤波算法判断图像的复杂程度，该图像为静态图像，保存第一帧为高清图像，其他重复的图像则进行图像替补标注，运用视频光流算法判断原始视频流运动复杂程度，连续视频帧为低密度运动视频，保存当前第一帧为高清图像，其他帧进行压缩并作低密度运动标识；对编码视频流进行解码，当检测到图像替补标识则读取保存的第一帧图像来显示，当检测到低密度运动标识，则获取当前低密度运动高清图像，通过视频补偿算法还原高清图像。利用图像替补视频压缩方法，既保证高清显示，也减少存储空间。 【EN】The invention discloses a video image-based ultralow code stream lossless compression method used in a K12 education stage, which comprises the following steps: capturing an original video stream; encoding an original video stream, judging the complexity of an image by using an edge filtering algorithm, wherein the image is a static image, storing a first frame as a high-definition image, carrying out image replacement marking on other repeated images, judging the motion complexity of the original video stream by using a video optical flow algorithm, storing a continuous video frame as a low-density motion video, storing the current first frame as the high-definition image, and compressing other frames and making a low-density motion identifier; and decoding the coded video stream, reading the stored first frame image for display when detecting the image replacement identifier, acquiring the current low-density motion high-definition image when detecting the low-density motion identifier, and restoring the high-definition image by a video compensation algorithm. By using the image replacement video compression method, high-definition display is ensured, and the storage space is reduced.

摘要：【中文】本发明公开了一种用于K12教育阶段基于人脸识别的图片自动合成方法，通过导入学生身份信息，生成学生唯一ID信息，将接送人员的身份信息与学生ID信息进行绑定，分别建立对应关系的接送人员以及学生的人脸图像模板。在同一时间内获取当前接送人员和学生的人像视频，显示在同一界面上,把获取到的当前接送人员和学生的人像视频流切成帧。每一帧人像视频通过人脸检测，提取当前接送人员和学生的人脸图像，分别与预先建立对应关系的接送人员以及学生的人脸图像模板进行匹配。匹配成功，则选定检测到接送人员以及学生现场的人像视频某一帧的图片作为合成照片的原始素材。根据需要的图片尺寸，把接送人员和学生原始素材自动合成为一种固定尺寸的照片。 【EN】The invention discloses an automatic picture synthesis method based on face recognition for a K12 education stage, which comprises the steps of generating unique student ID information by importing student identity information, binding the identity information of a receiving and sending person with the student ID information, and respectively establishing corresponding relations of the receiving and sending person and a face image template of a student. And acquiring the portrait videos of the current receiving and sending personnel and students in the same time, displaying the portrait videos on the same interface, and cutting the acquired portrait video streams of the current receiving and sending personnel and the students into frames. And extracting the face images of the current receiving and sending personnel and students from each frame of the face video through face detection, and respectively matching the face images with face image templates of the receiving and sending personnel and the students with a corresponding relationship established in advance. And if the matching is successful, selecting a picture of a certain frame of the detected person video of the receiving and sending personnel and the student scene as an original material of the synthesized photo. According to the required picture size, the original materials of the receiving and sending personnel and the students are automatically synthesized into a picture with a fixed size.

摘要：【中文】本发明公开了一种用于K12教育阶段的人脸识别校园区域预警方法及系统，包括：界定需要管控的区域，设置安全和预警范围；判断视频中出现的人员是否为内部人员以及对出现在视频中的人体行进方向进行分析，判断其是否有进入危险区域的趋势；对进入管控区域的人员进行甄别，提醒安保人员注意管控区域的异常状态。本发明通过采用区域识别的方式，基于人脸识别和行为分析，双重鉴别进入管控区域的人员，可明确鉴别人体和物体，并且进一步区分成人与幼儿、校外和校内人员，有效降低现有危险区域预警技术中的误报率，具有针对性的发出警报，提高预警准确性。 【EN】The invention discloses a campus area early warning method and a campus area early warning system for face recognition in a K12 education stage, wherein the campus area early warning method comprises the following steps: defining an area needing to be controlled, and setting a safety and early warning range; judging whether the personnel appearing in the video are internal personnel or not, analyzing the advancing direction of the human body appearing in the video, and judging whether the personnel have the tendency of entering a dangerous area or not; and screening the personnel entering the control area, and reminding security personnel to pay attention to the abnormal state of the control area. According to the invention, by adopting a region identification mode, based on face identification and behavior analysis, the personnel entering the control region are identified doubly, so that the human body and the object can be identified clearly, adults and infants, and personnel outside and inside the school can be further distinguished, the false alarm rate in the existing dangerous region early warning technology is effectively reduced, the alarm is sent out in a targeted manner, and the early warning accuracy is improved.

摘要：【中文】本发明公开了一种用于K12教育阶段基于人脸识别的接送系统，在应用端建立用户与对应人员关系，录入用户身份信息以及人像模板信息，并接收来自对应关系判断模块的对应关系匹配成功信息；人脸识别模块获取当前用户的人像特征，调取云服务器中的人像模板信息进行比对，若判断当前用户为特定用户，提取用户信息，发送至对应关系判断模块；对应关系判断模块根据当前用户的信息，检索与当前用户建立对应关系的人员身份信息，根据人员信息将预设信息发送至对应班级的信息接收模块，同时向闸机装置发出指令，允许用户通行，并将比对结果发送至用户应用端；班级信息接收模块接到信息后，在对应班级播放预设信息。 【EN】The invention discloses a receiving and sending system based on face recognition for K12 education, which establishes the relationship between a user and a corresponding person at an application end, inputs user identity information and portrait template information, and receives the information of successful matching of the corresponding relationship from a corresponding relationship judgment module; the face recognition module acquires the face features of the current user, calls face template information in the cloud server for comparison, extracts user information if the current user is judged to be a specific user, and sends the user information to the corresponding relation judgment module; the corresponding relation judging module retrieves the personnel identity information establishing the corresponding relation with the current user according to the information of the current user, sends preset information to the information receiving module corresponding to the class according to the personnel information, simultaneously sends an instruction to the gate device to allow the user to pass through, and sends a comparison result to the user application end; and after receiving the information, the class information receiving module plays the preset information at the corresponding class.

摘要：【中文】本发明公开一种用于K12教育阶段基于人脸识别的校园安全管理预警的方法，包括以下步骤：在终端建立人像模板并与服务器数据同步，摄像设备拍摄视频形成视频流上传至服务器，服务器从视频流中随机截取一帧图片与人像模板比对，若比对成功为允许进入人员则开门，若比对不成功则提醒进行访客登记，并设置访问时效和选择被访人员，将完成登记的人脸照片存储至访客登记分组中；访客可在有效期内通过人脸识别进入校园，访问时效到期后，访客信息自动归入历史登记人员组，访客访问权限到期后需重新登记。避免了繁琐的人工删除信息，减轻教职工工作量，同时可避免因人工操作而忘记删除导致来访人员多次入校，增加校园管理负担和安全风险的现象。 【EN】The invention discloses a campus safety management early warning method based on face recognition in a K12 education stage, which comprises the following steps: establishing a portrait template at a terminal, synchronizing the portrait template with data of a server, shooting a video by a camera device to form a video stream, uploading the video stream to the server, randomly intercepting a frame of picture from the video stream by the server to compare with the portrait template, opening a door if the comparison is successful, reminding a visitor to register if the comparison is unsuccessful, setting access time and selecting the visitor, and storing a face picture after the registration into a visitor registration group; the visitor can enter the campus through face recognition in the validity period, and after the visit aging is over, the visitor information is automatically put into a historical registration personnel group, and the visitor needs to be re-registered after the visit authority is over. The complex manual information deletion is avoided, the workload of teaching staff is reduced, and meanwhile the phenomena that due to the fact that the visitors forget to delete the information due to manual operation, the visitors enter the school for multiple times and the campus management burden and the safety risk are increased are avoided.

摘要：【中文】本发明公开了用于K12教育阶段基于人脸识别的多视频人员跟踪方法，建立行人的原始人像模板特征数据；对原始人像模板特征数据内已知人员的身份进行标注，并自动重建生成人脸特征数据模型；获取当前多监控视频流数据；对当前多监控视频流数据进行行人检测，获取行人的人脸特征数据；获取到的行人的人脸特征数据与自动重建生成人脸特征数据模型进行匹配；匹配成功，则提取当前行人的人脸特征数据；根据提取当前行人的人脸特征数据匹配行人出现的监控区域；根据监控区域的位置关系以及行人移动方向进行预判，在其他监控区域位置出现时，进一步提取该行人的人脸特征，并与之前提取的人脸特征进行融合，并继续跟踪。从而提高对跟踪对象进行跟踪的精准度。 【EN】The invention discloses a multi-video person tracking method based on face recognition for a K12 education stage, which is used for establishing original portrait template characteristic data of pedestrians; marking the identities of known persons in the original portrait template feature data, and automatically reconstructing to generate a facial feature data model; acquiring current multi-monitoring video stream data; carrying out pedestrian detection on the current multi-monitoring video stream data to obtain face feature data of pedestrians; matching the acquired face feature data of the pedestrian with a face feature data model generated by automatic reconstruction; if the matching is successful, extracting the face feature data of the current pedestrian; matching a monitoring area where the pedestrian appears according to the extracted face feature data of the current pedestrian; and pre-judging according to the position relation of the monitoring area and the moving direction of the pedestrian, further extracting the face features of the pedestrian when other monitoring area positions appear, fusing the face features with the face features extracted previously, and continuously tracking. Therefore, the tracking accuracy of the tracked object is improved.

摘要：【中文】本发明公开了一种用于K12教育阶段学校多人脸识别的方法，将人像模板库以及人员信息库存储于云服务器；拍摄区域内获取人像信息，区分符合条件的人体五官特征和人体活动特征；符合条件的人体五官特征则拍摄成照片，上传至本地服务器中；符合条件的人体活动特征则自动跟踪行人轨迹，对行人进行拍摄，并上传至本地服务器；本地服务器对照片或视频中检测到的一个或多个人脸进行提取，调取云服务器的人像模板库，与照片或视频中的人脸图像逐一进行比对，进行人脸锁定；将图片或视频归类至成功锁定的人员名下，并把预设的特效与非目标的人脸图像区域进行合成，遮挡无关人员的面部特征，推送至指定人员。 【EN】The invention discloses a method for multi-face recognition in a K12 education stage school, which comprises the steps of storing a face template library and a personnel information library in a cloud server; acquiring portrait information in a shooting area, and distinguishing human five-sense-organ characteristics and human activity characteristics which meet conditions; shooting the human five sense organ characteristics which meet the conditions into a picture, and uploading the picture to a local server; automatically tracking the pedestrian track according to the human body activity characteristics meeting the conditions, shooting the pedestrian, and uploading to a local server; the local server extracts one or more faces detected in the photos or videos, calls a face template library of the cloud server, compares the face template library with face images in the photos or videos one by one, and locks the faces; and classifying the picture or the video under the name of the person successfully locked, synthesizing the preset special effect and the non-target face image area, shielding the facial features of the irrelevant person, and pushing the facial features to the appointed person.

摘要：【中文】本发明公开了一种用于K12教育阶段基于人脸识别的校园行为分析方法，包括以下步骤：用户端建立幼儿与家长和老师一一对应的用户身份信息，并上传符合标准的幼儿人像图像，建立幼儿人像图像模板并存储至数据库中；将在学校区域内获取到的幼儿人像图像与数据库中幼儿人像图像模板比对，比对一致后把当前幼儿人体图像信息归类存储至服务器中；服务器通过归类的幼儿人体图像特征，分析识别幼儿表情和行为，并将表情和行为归类，统计出现次数最多的表情和行为，以此判断幼儿的心情状况和行为趋势，并将数据报告存储至数据库中,生成当日报告；同时提取幼儿历史心情和行为数据，累计当日报告，生成趋势报告，从而对幼儿的表情和行为进行有效监测分析。 【EN】The invention discloses a campus behavior analysis method based on face recognition for a K12 education stage, which comprises the following steps: the user side establishes user identity information corresponding to the infant, parents and teachers one by one, uploads infant portrait images meeting standards, establishes infant portrait image templates and stores the infant portrait image templates in a database; comparing the infant portrait image acquired in the school area with infant portrait image templates in a database, and classifying and storing the current infant portrait image information into a server after the comparison is consistent; the server analyzes and identifies the expressions and behaviors of the infant through the classified human body image features of the infant, classifies the expressions and behaviors, counts the expressions and behaviors with the most occurrence times, judges the mood and behavior trend of the infant according to the expressions and behaviors, stores the data report into a database, and generates a current day report; meanwhile, historical mood and behavior data of the infant are extracted, the current day report is accumulated, and a trend report is generated, so that the expression and behavior of the infant are effectively monitored and analyzed.