50results about How to "Implement face recognition" patented technology

The present invention discloses a face identification method based on a deep convolutional neural network. The method comprises: respectively feeding each image in a face identification database into three constructed deep convolutional neural network for feature extraction; performing normalization of output features, performing affine projection of the features to a low-dimensional space, obtaining a projection matrix, training a projection matrix through minimization of a ternary loss function, and obtaining feature vectors of each image; searching the weight value of each filter in the deep convolutional neural network through a gradient descending method, performing training test, and selecting a deep convolutional neural network having the highest average identification precision; and applying the selected deep convolutional neural network to a standard face identification database, performing Euclidean distance calculation of feature vectors of face images to be detected and each image, and if the feature vectors are smaller than a threshold value, determining that the images show the same person. Few images are used in the training, and the employed convolutional neural network is simple in structure so as to improve the face identification precision and reduce the training complexity.

The invention provides an information processing device. The information processing device comprises a storage unit and a data processing unit. The storage unit is used for receiving and storing dataand commands. The data processing unit is connected with the storage unit, and is used for receiving the data and the commands both sent by the storage unit, extracting and calculating key features included in the data and generating multi-dimensional vectors according to calculation results. The invention further provides an information processing method. The information processing device and theinformation processing method have the advantages that human faces are recognized accurately and rapidly; offline neural network operation is supported, and human face recognition and corresponding control works can be realized when user terminals / front ends are offline without assisted calculation of cloud servers; high portability is achieved, the information processing device and the information processing method can be applied to various application scenes and equipment, and design cost is saved greatly.

The invention discloses a intelligent business hall face recognition method based on lightweight deep learning, which comprises the following steps of: data preprocessing: detecting a face region in an acquired image, and then inputting the region into a face recognition network; design of a face recognition model; model training: learning and distinguishing features of a human face, gradually adjusting weight parameters of a network through multiple iterations based on a human face training data set subjected to data preprocessing, and finally obtaining a human face recognition model with good performance; model testing. According to the invention, a traditional business hall business handling mode is solved; through a face recognition mode based on deep learning, potential business recommendation can be actively provided for a user by analyzing user consumption information, precise marketing of customers is realized, business handling guidance is intelligently provided for the customers, and the user experience and the working efficiency of business handling personnel are greatly improved; under the condition of ensuring the model recognition precision, the model scale is greatlyreduced, and the face recognition of the mobile terminal is realized.

The invention discloses a face recognition method based on affine scale invariant feature transformation and structural similarity. The face recognition method relates to the technical field of face recognition. An operator AISFT (affine scale invariant feature transformation) is described by an image local characteristic to obtain a matching point between images; a redundant matching point is eliminated on the basis of a ''network leakage'' method; then, according to a structural similarity algorithm, the similarity based on a key point is obtained; and finally, the face recognition is realized. The face recognition method disclosed by the invention has the advantage of simplicity, easiness in realizing, correction and higher recognition rate.

The embodiment of the invention discloses a non-negative adaptive feature extraction-based human face identification method, a non-negative adaptive feature extraction-based human face identificationdevice, non-negative adaptive feature extraction-based human face identification equipment and a computer storage medium. The method comprises the following steps: integrating non-negative matrix decomposition, feature extraction and adaptive neighborhood learning into a uniform frame, acquiring local reconstruction data by a non-negative decomposition technology, minimizing reconstruction error in a non-negative reconstruction space and a feature embedding space at the same time, performing weight adaptive construction and label propagation learning on the reconstruction data, and performingminimizing learning by utilizing a projection-based feature approximate error item; alternately optimizing and learning a human face identification model to acquire an adaptive weight coefficient matrix for maintaining neighbor information, a projection matrix for extracting features and a neighbor maintaining non-negative decomposition matrix; and extracting the identification feature of a humanface test sample set by utilizing the projection matrix and realizing human face identification by utilizing the human face identification model and according to the identification features. Accordingto the technical scheme provided by the invention, the accurate rate of human face identification is improved.

The present invention provides a face detection and recognition method and device. The method comprises the steps of: pre-collecting at least one side face image, wherein the at least one side face image is a face image with a set angle of deflection; extracting corresponding side face features from each side face image; obtaining images to be detected, and extracting face angle features and facefeatures to be compared from the images to be detected; determining target angles of deflection corresponding to the images to be detected according to the face angle features; determining at least one side face image to be selected corresponding to the angles of deflection and the target angles of deflection from each side face image; perform comparison of the face features to be compared and theside face features corresponding to the side face images to be selected to obtain the similarity of the images to be detected and the side face images to be selected; and according to the similarity,recognizing a target side face images as a recognition result from the side face images to be selected. The face detection and recognition method and device can achieve face recognition based on theside face image.

The invention provides a face recognition method and system based on regional attention, a medium and a terminal. The method comprises the following steps: preprocessing a to-be-recognized face image to obtain a preprocessed full face image; according to a preset interception mode, intercepting the full face image, and obtaining an intercepted local face image; performing feature reasoning on the full face image and the local face image to obtain face features; performing feature matching on the face features to realize face recognition; the face recognition method based on the region attention mechanism is adopted, the overall features of the face image and the features of the local region are integrated, a scheme for learning by using different receptive field features is provided, feature attention of the non-shielded region of the face is enhanced through feature splicing and fusion, therefore, when the face of the user is shielded, the face recognition can still be effectively realized.

The invention provides a face recognition method and a mobile terminal. The method includes the steps of acquiring a target image; and performing face recognition on the target image by using a multi-domain hybrid network model, the multi-domain hybrid network model being a network model trained by using multi-domain samples, and the multi-domain samples being training samples sorted by pose classification. According to the invention, by the use of the multi-domain hybrid network model for face recognition, the multi-pose face recognition can be realized and the recognition effect of face recognition is improved.

The invention discloses a local binary model image describing method based on a filter group. The method comprises the following steps: dividing an inputted facial image into blocks for a plurality of image blocks; carrying out convolution filtering to each image block based on the filter group; performing LBP coding to each image block after the process of convolution filtering so as to obtain a plurality of FBLBP values for each image block; making statistics to the obtained plurality of FBLBP values for each image block; obtaining the FBLBP histogram features in each filter mode and combining the FBLBP histogram features in all filter modes in the filter group sequentially as the FBLBP histogram features of the image block; and sequentially extracting the FBLBP histogram features of each image block; concatenating the FBLEP histogram features to develop the histogram feature vectors of the facial image. According to the invention, an image is coded so that multi-scale information of the image can be obtained, that the LBP operators are more robust and immune to interference, and that characterization of LBP descriptors to the image is enhanced.

The invention discloses a face recognition temperature measurement campus access control system based on an embedded GPU. The system comprises a school personnel information background management unit, a face detection and recognition unit, a face recognition display interface unit, an embedded GPU parallel acceleration unit, an infrared temperature measurement camera device, a display device, an alarm access control linkage device and the like. The school personnel information background management records personal information of students, the alarm access control linkage device is responsible for distinguishing the students with detected high temperature and reporting and managing the students, and the face detection and recognition unit mainly realizes face detection, face recognition, mask recognition, living body detection and the like. The face recognition display interface unit processes and displays face information captured by the infrared temperature measurement camera device in real time. A GPU parallel acceleration unit is adopted to carry out acceleration calculation processing on a face detection and recognition algorithm, and the real-time performance of the system is improved.

The embodiment of the invention provides a face recognition method and device. In the embodiment, the face recognition method comprises the steps that a face recognition request which is received manytimes is subjected to data encapsulation, and a task data packet is obtained; the task data packet is cached to a task stack for waiting processing; when any thread in a thread pool is idle, the taskdata packet in the stack is obtained; based on a preset neural network model, the data of a plurality of images in the task data packet is subjected to face recognition, and face recognition resultsof the data of the images are obtained.

The embodiment of the invention discloses unlocking control methods and related products. One of the methods includes: acquiring a face image; extracting a rough feature set and a fine feature set from the face images; carrying out verification operation on the rough feature set and the fine feature set; and executing a next-step unlocking process when verification of the rough feature set and thefine feature set is successful. According to the embodiment of the invention, both the rough feature set and the fine feature set extracted from the face image can be verified, and the next-step unlocking process is executed only when verification of the both is successful. Therefore, the method can be used to realize face recognition, and improves the security of a mobile terminal.

The invention discloses an image processing method for protecting privacy. The method comprises the following steps: S1, acquiring image information; S2, converting the image information into an image data matrix; S3, performing convolution operation on the image data matrix and a group of mask matrices to obtain a group of new data matrices; and S4, re-coding the obtained new data matrices into a group of new arrays. Compared with the prior art, the image processing method for protecting privacy provided by the invention has the advantages that a series of mask matrices are designed, and the original image matrix and the mask matrices are subjected to convolution operation, so that a series of new matrices are obtained; in the newly generated matrices, the visual information of the original image is recombined, and the original coding mode is destroyed, so that an observer cannot directly obtain the visual information, and it is ensured that privacy cannot be leaked from the source.

The invention provides a face recognition method and a face recognition device. The method comprises the following steps: acquiring reflected signals of a face of a target recognition object, wherein the reflected signals comprise an optical signal and an ultrasonic signal; and performing feature extraction on the reflected signals to generate actual feature data of the face; comparing the actual feature data with ideal feature data in a pre-established feature database, and determining identity information of the target recognition object, wherein the feature database comprises identity information of a specific user and ideal feature data corresponding to the identity information, and the ideal feature data comprise optical feature data and ultrasonic feature data corresponding to the face of the specific user. According to the invention, the optical signal and the ultrasonic signal are combined together, the feature data of the target recognition object can be extracted through the ultrasonic signal under strong light exposure under the condition that the optical signal is weak, all-weather face recognition is achieved, and the environmental adaptability of a face recognition mode is improved.

The invention discloses a child behavior track positioning method and system, and the method comprises the steps: constructing a model library, and carrying out the feature extraction of an image in the model library, and obtaining a first feature library; performing face detection and human body detection on the collected video to generate a face detection frame and a human body detection frame; performing first matching on the human face detection frame and the first feature library, performing feature extraction on the human face detection frame based on a first matching result to obtain a second feature library, performing human face tracking matching on the to-be-recognized object successfully matched for the first time, and performing second matching based on the second feature library based on a human face tracking matching result; recording a coordinate position of successful first matching or successful face tracking matching or successful second matching in the image sequence so as to generate a track; through comprehensive face and human body information recognition, the problems of face shielding and back-to-back camera of the body are solved, and the behavior track of the child in the classroom is accurately tracked.

The embodiment of the invention relates to the field of artificial intelligence, and discloses a face recognition method and device, electronic equipment and a storage medium. The method comprises thesteps of inputting a to-be-recognized face image into a pre-trained convolutional neural network model to obtain a recognition result output by the convolutional neural network model, wherein the convolutional neural network comprises an affine estimation branch network and a convolutional integral branch network, the affine estimation branch network is used for face detection and face alignment,the convolution branch network is used for carrying out identity recognition on the face image. An affine estimation branch network is added to a convolutional neural network model and used for achieving face detection and face alignment, face detection, face alignment and face recognition are further completed in the convolutional neural network model at the same time, and therefore end-to-end face recognition is achieved.

The invention provides a system for assisting in catching a criminal by using an elevator. The system comprises a single chip microcomputer, a first camera, a memorizer and a network connection module; the first camera is used for shooting face images of passengers in a lift car, face image information of criminals is saved in the memorizer, and the single chip microcomputer is used for performingface recognition, sending alarm information to a public security system timely and guiding an elevator control system in being locked. A method for assisting in catching the criminal by using the elevator particularly comprises the steps that A, face recognition is performed on the passengers in the lift car; B, whether or not the criminal exists among the passengers is judged; C, whether or notgeneral people exist among the passengers; D, the elevator control system is locked; E, police commissioners arrive at the scene to catch the criminal. The system and method for assisting in catchingthe criminal by using the elevator has the advantages that personal identity verification can be performed on the passengers, and the criminal can be stuck in the elevator in appropriate conditions, thereby providing convenience for the public security to catch the criminal.

The invention provides a millimeter wave image-based face recognition method and system. The method comprises the steps of performing size processing on a millimeter wave image to obtain a processed image consistent with a registered image in a face image retrieval library in size; extracting face features of the processed image to establish feature extraction mapping; and comparing the feature extraction map with a pre-established preliminary feature map of the registered image to obtain a matched preliminary feature map. According to the invention, face recognition can be realized based on the millimeter wave image generated by the millimeter wave security check imaging system.

The invention relates to the field of face recognition, and discloses a face recognition device capable of intelligently lifting according to the height. The face recognition device comprises a wall,wherein a main box body cavity penetrating from left to right is formed in the wall; a main box body is fixedly connected to the front and rear end walls of the main box body cavity; a main lifting cavity with a rightward opening is formed in the main box body; a main motor fixedly connected with the main box body is arranged at the upper side of the main lifting cavity; and a main screw rod extending downwards into the main lifting cavity is fixedly connected to the lower end surface of the main motor. The height of a user is detected through an infrared sensor, a camera is automatically switched between a high-position overlooking mode and a low-position head-up mode, and face recognition for the user in two height ranges can be achieved, so that the tedious step that the height of the camera needs to be adjusted for each recognition is conducted is avoided, and the recognition efficiency is improved; and meanwhile, through automatic contraction in the lifting process of the camera,the camera is prevented from colliding with the user, so that potential safety hazards are reduced.

The invention discloses a photo live broadcast system with a face recognition function and a method thereof. The system comprises a server, a cloud storage system, a playing terminal and a face recognition system; the server is used for sending the first photo to the cloud storage system for storage, and inquiring the face recognition feature information from the first database and sending the face recognition feature information to the playing terminal; the face recognition system is used for responding to a face recognition processing request sent by the server and returning the first picture and the second face recognition feature to the server; and the playing terminal is used for uploading the second photo to the server and querying an access request of the first photo with the same face recognition characteristic as the second photo, and displaying the face recognition characteristic, the face image and / or the first photo according to the query result. According to the invention,immediate uploading and immediate browsing of mass photos are realized, the problems caused by high concurrency, high access and high downloading are solved, and immediate face recognition of the photos is realized, so that photo live broadcasting is realized.

The invention relates to an operation platform for intelligently customizing a bus based on a mobile internet and an operation method thereof. The operation platform comprises a consulting and ticketbooking module, a road transportation administration department reporting module, a vehicle scheduling module, a passenger ticket checking and luggage security checking module, a driving record monitoring module and an emergency automatic alarm module. The operation method comprises the following steps: collecting vehicle demand information and customizing a demand scheme; determining a travel, and distributing travel information; reserving seats, and deploying vehicle types according to the number of people; carrying out ticket checking, luggage security checking and getting-on guiding; monitoring the travel; and carrying out service evaluation. The operation platform for intelligently customizing the bus based on the mobile Internet is used for collecting, emphasizing and analyzing the user demand according to the bus use requirements, and the travel route is planned in a real-time and diversified manner according to the route demand of the client, so that the time cost of the bus traffic route is greatly reduced, the labor cost is optimized, and the optimized bus traffic route can meet the travel demand of the user.

The invention discloses a medication checking device, which is technically characterized by comprising a face recognition device and a supporting table, and further comprising a standing supporting assembly which is arranged on the supporting table and used for a medication nurse to stand and used for heightening so as to conveniently check the identity; a supporting pipe barrel is fixedly installed on the supporting table through a connecting assembly, and a supporting rod is connected into the supporting pipe barrel in a sliding mode in the vertical direction; a height driving assembly is used for driving and controlling the supporting rod to vertically slide; an angle driving assembly is used for driving and controlling the angle adjustment of the face recognition device; an angle locking assembly is used for locking the angle of the face recognition device; a fixing assembly is used for mounting the face recognition device on the angle driving and controlling assembly; face recognition can be achieved through the face recognition device, the device is more convenient to use, and the height and the angle can be conveniently adjusted.

The invention provides a payment method and cloud platform system based on big data and artificial intelligence, the payment cloud platform system based on big data and artificial intelligence comprises a first mobile terminal and a server, and the method comprises the following steps: the first mobile terminal of a payment initiating user responds to a common payment operation initiated for a to-be-paid bill; photos including at least two common payment users are acquired, and the photos are sent to a server; the server performs face recognition on the received photo and determines identity information of the common payment users; the server splits the to-be-paid bill according to the number of the common payment users to obtain a plurality of sub-bills; and the server pushes payment requests corresponding to the sub-bills to a second mobile terminal of the common payment user according to the identity information of the common payment user. Therefore, face recognition can be carried out on the group photo, so that multi-person sharing payment or bill sharing payment bill pushing can be automatically realized, and the convenience and efficiency of payment actions can be improved.

The invention discloses a charging station maintenance method, device and equipment and a computer readable medium, and belongs to the technical field of charging station maintenance. The method comprises the following steps: acquiring a video of a charging station and an equipment state of the charging station; according to the video of the charging station, judging whether an abnormal behavior exists in the charging station or not; judging whether abnormal equipment exists in the charging station or not according to the equipment state; and if the charging station video has the abnormal behavior and / or the abnormal equipment, obtaining a maintenance process to start charging station maintenance. According to the technical scheme disclosed by the embodiment of the invention, multiple functions of video playing, abnormity identification, intelligent detection and efficient operation and maintenance can be integrated, and timely and comprehensive maintenance and management of the charging station can be realized.

The invention discloses a mask face living body detection method based on a support vector machine, and the method comprises the steps: carrying out the positioning of face key points through a face key point positioning algorithm based on a multi-stage cascade regression tree, and automatically generating mask face samples in batches according to the positioning points; constructing a face library without a mask and a face library with a mask; performing face recognition; extracting an LBP feature of the face image, and converting an LBP feature matrix into a histogram vector as a texture feature vector for living body detection; training and storing a support vector machine model based on texture features; carrying out living body detection; and outputting a verification result according to the face recognition result and the living body detection result. The face key point model is trained, key points of the nose bridge and the chin of the face can be positioned, and whether the face wears a mask or not is judged by detecting whether the key points are shielded or not; according to the method, the support vector machine is trained, a rapid in-vivo detection function is realized, the function is simultaneously suitable for an ordinary face and a face with a mask, the rapidity and accuracy of face recognition are not influenced, and the safety of the system is improved.

The invention is suitable for the technical field of image processing, and provides a face recognition method, a face recognition device and terminal equipment, and the method comprises the steps: obtaining a visible light image and a near-infrared image; performing face feature extraction on the visible light image and the near-infrared image to obtain visible light face feature information and near-infrared face feature information; performing face recognition based on the visible light face feature information, if a matched visible light face image sample exists, determining a registered user associated with the matched visible light face image sample as a face recognition result, and otherwise, performing face recognition based on the near-infrared face feature information; and if thematched near-infrared face image sample exists, determining a registered user associated with the matched near-infrared face image sample as a face recognition result, and otherwise, outputting a reminding message. By means of the method, face recognition can be conducted on the global race with various skin colors.

The present invention provides a face recognition method and device, wherein the method includes: acquiring the reflection signal of the face of the target recognition object; wherein the reflection signal includes an optical signal and an ultrasonic signal; performing feature extraction processing on the reflection signal to generate a face The actual feature data; compare the actual feature data with the ideal feature data in the pre-established feature database to confirm the identity information of the target identification object; wherein, the feature database includes the identity information of the specified user and the ideal feature corresponding to the identity information data; the ideal feature data includes optical feature data and ultrasonic feature data corresponding to the specified user's face. By combining the optical signal and the ultrasonic signal, the present invention can extract the characteristic data of the target identification object through the ultrasonic signal when the optical signal is weak when the strong light is irradiated, so as to realize all-weather face recognition and improve the face recognition. The environmental adaptability of the identification method.

The invention discloses an adaptive local face recognition method and system used in the field of security protection. The method comprises the steps: firstly determining the region type of a detected face region, and then carrying out the feature similarity calculation of the same region type. The system comprises an image acquisition module, a local face detection module, a face detection post-processing module and a face recognition module. For the problem of face posture diversity in the field of security protection, face recognition can be realized under the condition that a front-view camera is not needed. According to the method, a large number of samples which are locally visible and worn with masks or other shelters are used in model training, so that the situation that the faces with worn masks cannot be recognized in the gates of airports and stations can be greatly improved, and the accuracy and efficiency of face recognition in the security field are greatly improved.

The invention discloses a door opening system through a dry contact 485 protocol. The door opening system comprises a door opening terminal and an adapter module, wherein the door opening terminal isin bidirectional connection with the adapter module, the adapter module comprises a dry contact input wiring port and an output wiring port, the dry contact input wiring port is connected with an output line of the door opening terminal, and the output wiring port is connected with an input box of the door opening terminal. According to the door opening system through the dry contact 485 protocol,the dry contact input wiring port of the adapter module is connected with theoutput line of the door opening terminaland the output wiring port of the adapter module is connected with the input box of the door opening terminal, after the module receives the signal, an instruction is output to the device through the adapter module, the device executes a corresponding action according to a hardwareprotocol to open the door at the cloud end, face-scanning door opening can be achieved in an off-network state, face recognition local door opening can be achieved, and the response speed is greatlyincreased compared with that of cloud-end door opening.