56results about How to "Improve face recognition rate" patented technology

The invention provides a face recognition method based on a convolutional neural network. The face recognition method comprises the following steps: carrying out necessary pretreatment at early stage on a facial image to obtain an ideal facial image; selecting the ideal facial image as the input of the convolutional neural network to enter U0, wherein the output of the U0 enters UG, and the output of the UG is taken as the input of US1; extracting edge components in different directions in an input image as first-time feature extraction and outputting to the input of special UC1 through supervised training by the S nerve cell of the US1; taking the output of the UC1 as the input of US2, completing the second-time feature extraction by the US2 and taking as the input of UC2; taking the output of the UC2 as the input of US3, and completing the third-time feature extraction by the US3 and taking as the input of UC3; taking the output of the UC3 as the input of US4, and obtaining the weight, threshold value and neuron plane number in each layer in a supervision competitive learning mode by the US4 and taking as the input of the UC4; taking the UC4 as the output layer of the network, and outputting the final mode recognition result of the network determined by the maximum output result of the US4. According to the face recognition method, the recognition rate of faces in complex scenarios can be improved.

The invention relates to a method for performing face recognition by combining the rarefaction of shape characteristic, belonging to the image processing field. The method comprises the following steps: performing textural characteristic extraction and shape characteristic extraction to all the facial images in a training set based on constrained sampling, obtaining textural characteristic matrix and shape characteristic matrix, corresponding one type of the textural characteristic matrix and shape characteristic matrix to a plurality of face images of one person in the training set respectively; performing textural characteristic extraction and shape characteristic extraction to the face image of a person to be identified based on constrained sampling, obtaining the textural characteristic vector of a image to be identified; calculating the textural residual error and shape residual error of each type in the training set; using the linear coefficient of the shape characteristic vector of the training set to represent the shape characteristic vector of the image to be identified; and using the type of the training set with the maximum comprehensive similarity to the face to be identified as the identifying result of the person to be identified. The method of the invention has higher face recognition.

The invention discloses a method for checking class attendance on the basis of multi-face data acquisition strategies and deep learning. By the aid of the method, the technical problem of low recognition rates of existing methods for checking attendance on the basis of face recognition can be solved. The technical scheme includes that multiple objects are detected and extracted by the aid of AdaBoost algorithms and skin color models. Only a piece of video needs to be shot on every face participating in attendance checking at one step, faces in video sequences are detected and extracted, and face databases can be completely created. The method has the advantages that learning can be carried out on face features in the face databases in different scenes by the aid of simplified LeNet-5 models on the basis of depth convolutional neural network LeNet-5 models by the aid of processes for recognizing the faces on the basis of deep learning, and novel features can be represented by means of multilayer nonlinear transformation; intra-class change of illumination, noise, attitude, expression and the like is removed from the novel features as much as possible, inter-class change generated by identity difference is reserved, and accordingly the face recognition rates of the processes for recognizing the faces in practical complicated scenes can be increased.

The invention, which relates to the technical fields including pattern identification, image processing and computer vision and the like, provides a face identification method based on a wavelet multi-scale analysis and a local binary pattern (LTP). The method comprises the following steps: selecting an appropriate face image; carrying out a multi-scale wavelet analysis on a training image to obtain a first-level low frequency approximation image and a second-level low frequency approximation image; utilizing an LTP algorithm to carry out conversion on the low frequency approximation images to obtain LTP characteristic values of all pixel points; carrying out statistics on LTP histograms of the images by utilizing a blocking method and connecting the blocked histograms of the images of the two levels to obtain characteristic vector representation of the face image; and for a to-be-identified face, obtaining a characteristic vector of the to-be-identified face image and then using X <2> probability statistics to complete face identification. According to the method provided by the invention, an image noise effect can be effectively reduced; extraction capability for image texture characteristics can be enhanced; besides, the method has advantages of good robustness, high identification rate, fast calculating speed and important practical value.

The invention discloses a face recognition method based on a face image feature extreme learning machine. The method comprises steps as follows: preprocessing original images; performing principal component analysis on samples to obtain feature face spectrums, and projecting the images into a feature domain; then establishing the mapping relation between face images and face tags with an extreme learning machine algorithm; finally, deducting tag attributes input into the face images by using the extreme learning machine. According to the method, the advantages of the extreme learning machine are taken, the complexity of estimation and optimization of parameters of a traditional neural network is decreased, the training time is further shortened, and the recognition rate of the face images is further increased.

The invention discloses a face recognition method for a bottom-occluded face image. Feature images are extracted for a to-be-recognized face image with bottom occlusion and a plurality of non-occludedface image samples respectively. According to the extraction method, a face region image is obtained through extraction and input into a convolutional neural network, feature maps output by the lastconvolutional layer are used as initial feature maps of the face image, each initial feature map is multiplied with a weight matrix to obtain one feature image, the values of weight elements corresponding to the upper portions of the initial feature maps in the weight matrix are large, the values of the weight elements corresponding to the lower portions of the initial feature maps are small, andthen the similarity between the feature image of the to-be-recognized face image and the feature image of each face image sample is calculated to complete face recognition. According to the method, byperforming region weight processing on the feature maps of the convolutional neural network, non-occlusion features are enhanced, and the face recognition rate of the bottom-occluded face image is increased.

The invention provides a method for improving the human face identification rate. The method is suitable for a digital camera with a memory unit. Human face information is stored in the memory unit, and includes human face characteristics. The method for improving the human face identification rate comprises the following steps: retrieving a first human face image of a target human; executing a human face identification program to determine whether the first human face image is corresponding to human face information in the memory unit or not through the human face identification program; and executing an updating program when the first human face image is corresponding to the human face information so as to increase the human face characteristics corresponding to the target human in the memory unit.

The invention relates to a face identification method based on the manifold learning, which comprises the following steps: offering partial category relation saving the face data from the data partial structure; converting the optimum partial of every point to the total optimum with the arrangement technique; getting the data projection matrix from higher dimensional to lower dimensional by proceeding characterized dissociation for arrangement matrix, the original training sample and transposed product; mapping the identifying face image to the lower dimensional space; classifying with the nearest neighbor method in the lower dimensional space. The invention provides the higher face identification ratio, which provides higher identification ratio than the identification with principal component analytical method.

The invention provides an apparatus and method for registering plurality of facial images for face recognition. A facial image registration apparatus and method are provided for registering a plurality of facial images to be used for face recognition. A plurality of facial images are detected and sorted from a plurality of captured images input from a camera unit at the request of a user. A face feature value is calculated for each of the facial images. The face feature value is a score that is applied according to face features of each of the facial images. Facial images selected according to the calculated face feature values are registered as target facial images to be used for face recognition. As such, various facial images may be registered at a user's single request, thus improving a face recognition rate.

The invention provides a face recognition method and device, and relates to the technical fields of pattern recognition and biological feature recognition. The face recognition method and device can be used for carrying out effective recognition under various illumination conditions, and enhances the performance of face recognition. The method comprises the following steps: acquiring an original face test sample to be recognized, and carrying out Gaussian difference filter treatment on the original face test sample to obtain a filtered face test sample; comparing the original face test samplewith a prestored original face training sample set registered under normal illumination conditions, comparing the filtered face test sample with a prestored face training sample set subjected to Gaussian difference filter, and finding out the recognition object corresponding to the original face test sample from the original face training sample set; and computing to obtain an overall SCI (Sparsity Concentration Index) of the original face training sample of the recognition object and the filtered face training sample, and judging whether the original face test sample is a registered face according to the SCI.

The invention provides a non-adjacent graph structure sparse face recognizing method. The method comprises the steps of enabling the non-adjacent graph structure to be sparse; searching by blocks or combination; measuring the structure sparseness; performing structure sparse reconstruction. According to the method, the system performances are improved through non-adjacent graph structure sparseness based on an SRC model; the blocking of the non-adjacent graph structure is dynamically performed by an overlapping manner and cannot be predicated; the components can be adjacent or non-adjacent; all possible combinations can be searched by the combination method to gain adjacent or non-adjacent blockings, so as to achieve the non-adjacent graph structure sparseness; to avoid combination explosion in search, the method of searching by blocks or combination is provided for limiting the search space as well as generating computer acceptable basis subset space; the non-adjacent graph structure sparse reconstruction is carried out by the structure greedy algorithm; when in iterating of the algorithm, the base blocks are selected according to the contribution degree of the base blocks; the non-adjacent graph structure sparseness is measured according to the coding complexity. With the adoption of the method, the face recognizing rate can be obviously increased.

The embodiment of the invention provides an automatic near-infrared face registration method and device, an automatic near-infrared face recognition method and device and a storage medium. The automatic near-infrared face registration method comprises the steps that S210, a face near-infrared image which contains a target face and is automatically collected by a near-infrared camera while an RGB camera collects a face RGB image containing the target face is automatically acquired; S220, a target near-infrared image containing the target face is obtained at least based on the face near-infraredimage; S230, target object information relevant to the target face in the face RGB image is acquired, wherein the target object information is identification information of an object having the target face; S240, near-infrared registration information corresponding to the target face is obtained at least based on the target near-infrared image and the target object information; and S250, the near-infrared registration information is stored into a near-infrared base library. In this way, the requirement on user cooperation is low, user experience is good, and the face recognition rate is high.

The present invention discloses a selective integrated human face recognition method using a genetic algorithm fused with differential evolution. The method comprises: firstly, extracting an HOG feature of a human face image; then, performing dimension reduction on the human face image by using a PCA algorithm to reduce calculation complexity; and finally, performing classified recognition by using data after dimension reduction and applying a GADESEN algorithm. According to the method, with a support vector machine used as a base classifier, N samples are extracted with replacement from an original training set, and according to the method, iteration is performed for T times, and a sample set generated each time is used to train a base classifier model, N generated base classifiers are real-coded to generate an initial population, and the difference vector is used to guide the mutation to produce the high quality individual. The crossover operation uses the parent individual and the mutated individual to produce the cross individual, which increases the individual's diversity, Retention strategy for genetic evolution.

The invention provides a face recognition method and system. The method has the advantages that the computing speed is high, the computing result approaches the optimal solution, and the robustness ishigh; a training sample is aligned to a testing sample, and the face recognition rate is increased; the generated aligned face image and an original face image are simultaneously adopted as the training sample, and a training sample set composed of all samples is effectively expanded; meanwhile, face edge characteristics are extracted as a novel training sample for classified calculation, and theinfluence caused by factors such as complex background or illumination unevenness is effectively lowered; a residual error obtained under the two training samples is effectively combined, and the final face recognition rate is increased finally.

The invention relates to a face recognition method based on an uncertainty quantization probability convolutional neural network. The method comprises a training stage and an identification stage. Theprobabilistic convolutional neural network is trained through samples with known categories. The extraction of face features is realized by a learning process of a probability convolutional neural network. The description of the face features is represented by the connection weight; testing the trained probability convolutional neural network by using a training sample, and determining a classification threshold. A to-be-identified sample is input into the probability convolutional neural network. An output vector of the probability convolutional neural network is calculated. A maximum component is compared with a classification threshold value to give an identification result. Uncertainty quantitative analysis is made on the identification result, and a mean value and variance estimationof the identification result are provided. The probability convolutional neural network system is adopted. The explicit feature extraction process is avoided through the difference extraction layer and the feature mapping layer. The feature which greatly contributes to the construction of the sample space is implicitly obtained from the sample. Compared with the prior art, the recognition rate and the anti-interference performance are higher.

The invention provides a face identification method based on multi-sample expansion cooperation expression classification; the method comprises the following steps: creating a mirror image face imageon the image mirror surface property basis; preparing two random original samples and mirror image samples of the same type, and using smooth median samples of said samples so as to build a novel virtual sample; using the European style distance to select training samples approaching to test samples. The method can carry out parameter weight fusing for training samples formed in different paths, can identify faces according to a classification method based on cooperation expressions, can create various virtual training samples, and can simplify the operation complexity, thus improving the faceidentification effect.

The invention discloses a face image quality evaluation method, device and a computer readable storage medium. According to the face image quality evaluation method, the average brightness, the noiseintensity and the feature definition of the face image are obtained, the brightness evaluation value, the noise evaluation value and the feature definition evaluation value of the face image are obtained through calculation, and finally the quality of the face image is evaluated according to the evaluation values of the brightness, the noise and the feature definition of the face image. The invention also discloses a face image quality evaluation device and a computer readable storage medium, which are used for realizing each step of the method. According to the method, the device and the computer readable storage medium, the quality of the face image is comprehensively evaluated from three dimensions of brightness, noise and feature definition, the difference of the quality of the face image caused by subjective difference of human eyes is avoided, and the face recognition rate is increased.

The embodiment of the invention provides a face recognition method and device. The method comprises the following steps: acquiring a first face image; determining a target makeup degree parameter of the first face image; determining a target restoration parameter corresponding to the target makeup degree parameter, and performing image processing on the first face image according to the target restoration parameter to obtain a second face image; matching the second face image with a preset face template; when the second face image is successfully matched with the preset face template, confirming that the first face image is successfully recognized. Through the embodiment of the invention, the makeup face image can be restored, and the face recognition rate is improved.

The invention provides a three-dimensional face representation method and a parameter measurement device and method thereof, the device comprises a data acquisition module set, a light source array group module, a support and a processing terminal, the data acquisition module set is connected with the processing terminal, and the light source array group module is arranged on the support; the data acquisition module group is used for measuring and reconstructing the geometrical parameters of the human face of the measured object and the reflection parameters of the skin surface, the data acquisition module group comprises a pair of data acquisition modules, and the data acquisition modules are arranged in a bilateral symmetry mode and are both arranged to form a first preset angle with the plane which is perpendicular to the ground and intersects with the support; the light source array group module is used for generating mirror reflection light in each area of the skin surface of the measured object; and the processing terminal is used for receiving the data which is transmitted from the data acquisition module and is related to the measurement of the geometrical parameters and the reflection parameters, processing and reconstructing the measured object, and can construct a face database which is beneficial to improving the face recognition rate.

The invention relates to a face recognition algorithm and a face recognition device adaptive to an illumination interference environment, the face recognition algorithm comprises a step of establishing a face data vector set and a step of carrying out face recognition; in the step of face recognition, detecting a to-be-recognized portrait photo by using an MTCNN, respectively inputting the obtained face images into a FaceNet model and an LBP model to obtain corresponding vectors, then calculating the Euclidean distance between the vector and each vector in a face data vector set, and then performing weighted summation calculation according to the Euclidean distance to obtain a face image; and taking the training portrait photo corresponding to the minimum value of the weighted summation as a face recognition result of the portrait photo to be recognized. According to the method and device, the weights of the two algorithms are adaptively adjusted, and adaptive environment illumination is achieved, so that the illumination robustness of the neural network during face recognition is improved, and the face recognition rate in an illumination interference environment is improved.

The invention provides a Hog and Gabor feature efficient fusion fast face recognition method based on near-infrared face image, belonging to the technical field of pattern recognition and image processing. The method includes Hog feature extraction and Gabor feature extraction for near-infrared face image samples respectively. The non-negative matrix decomposition method is used to reduce the dimension of two kinds of eigenmatrices respectively. The Hog feature and Gabor feature are fused to obtain the fused feature. The linear discriminant analysis (LDA) was used to reduce the dimension of the fusion feature matrix, and the feature vectors of the training samples after the transformation matrix of the second projection and the second reduction are obtained. The nearest neighbor algorithmKNN is used to classify and recognize the test samples based on the distribution of training samples. The invention ensures the comprehensiveness of acquiring characteristic information, effectively improves the characteristic representation efficiency, reduces the cost of algorithm operation storage and time consumption, and improves the efficiency of face recognition process.

The invention discloses a self-depth optimization face recognition device and a method thereof. The device comprises a shell and a standing plate located at the bottom end of the shell. A sliding groove is formed in the side, close to the standing plate, of the shell. A sliding block matched with the sliding groove is arranged in the sliding groove. A first motor is arranged at the bottom end in the shell, a lead screw extending into the sliding groove and penetrating through the sliding block is arranged at the output end of the first motor, a three-dimensional face recognition mechanism is arranged in the sliding block, and a distance sensor matched with the three-dimensional face recognition mechanism is arranged on one side of the shell. The beneficial effects are that the motor drivesthe screw rod to rotate to lift the three-dimensional face recognition mechanism to a proper height, so that the three-dimensional face recognition mechanism is suitable for people of different heights; the internal face recognition reference library can be dynamically updated based on a self-depth optimization algorithm according to a time period and a feature data change rate. Therefore, the face recognition rate of people whose faces are in a rapid growth and development or aging stage is greatly improved, and the application range and depth of the face recognition technology in differentpeople and different fields are greatly improved.

The invention relates to a sparse representation face identification method based on constrained sampling and shape features, belonging to the field of image processing. The method comprises the following steps: extracting the textural features and shape features of face images in a training set based on constrained regions to obtain a textural feature matrix and a shape feature matrix, wherein each category of the textural feature matrix and the shape feature matrix respectively corresponds to the multiple face images of one person in the training set; and comparing each category in the training set with the obtained textural feature vector and shape feature vector of the face image to be identified, and taking the category in the training set corresponding to the maximum comprehensive similarity value as the identification result of the person to be identified. The method has higher face identification rate.

The invention discloses an application system of a face recognition technology based on a bus scene in the aspect of public safety. The system comprises a blacklist face library of criminal suspects wanted by the public security system, a monitoring camera module, a face recognition module and an automatic alarm module; the monitoring camera module comprises a front door camera module and a rear door camera module; the application system is high in recognition degree, automatic alarm can be achieved, and the safety environment on a bus is improved.

The invention discloses a face recognition method based on structuralized factor analysis. The problem that the face recognition accuracy rate is low due to the fact that local clustering features and a global distribution structure of face image data cannot be kept in the prior art is mainly solved. The face recognition method comprises the achieving steps of 1 dividing a face image data set; 2 performing clustering analysis on all training sample data; 3 calculating an optimal feature projection matrix through Gibbs sampling; 4 extracting low-dimension features of all test sample data and the training sample data; 5 recognizing a face image. The face recognition method synthesizes the local clustering features and the global distribution structure of the face image data and improves the face recognition accuracy rate.

The invention discloses a night face recognition optimization method. The method comprises the following steps: detecting a face, and detecting the call data of a camera; when the human face is detected and the camera is detected to be called, storing the current screen brightness parameter, and defining the screen brightness corresponding to the current screen brightness parameter as the originalscreen brightness; increasing the original screen brightness to be at least the screen brightness corresponding to the set parameter; collecting the face image, judging whether the face image passesthe authentication, and if so, adjusting the improved screen brightness to the original screen brightness based on the current screen brightness parameter; otherwise, prompting that the face recognition fails. By presetting the set parameters, night face recognition is better achieved under the screen brightness corresponding to the set parameters, the current screen brightness parameters are compared with the set parameters, the current screen brightness is adjusted, and the face recognition rate is increased.

The invention discloses a face recognition method for face images with lower occlusions. For face images to be recognized with lower occlusions and several face image samples without occlusions, feature images are extracted respectively. The extraction method is as follows: extracting and obtaining The image of the face area is input into the convolutional neural network, and the feature map output by the last layer of convolution layer is used as the initial feature map of the face image, and each initial feature map is multiplied by the weight matrix to obtain the feature image. The value of the weight element corresponding to the upper part of the initial feature map in the matrix is ​​larger, and the value of the weight element corresponding to the lower part of the initial feature map is smaller, and then calculate the feature image of the face image to be recognized and the weight of each face image sample The similarity of feature images completes face recognition. In the present invention, the area weight processing is performed on the feature map of the convolutional neural network, so as to strengthen the unoccluded feature and improve the face recognition rate of the lower occluded face image.

The invention discloses a selective integrated face recognition method based on genetic algorithm fusion differential evolution. Firstly, the HOG feature of the face image is extracted, and then the PCA algorithm is used to reduce the dimension of the face image, so as to reduce the computational complexity. Finally, use After dimensionality reduction, the GADESEN algorithm is used to classify and identify the data. This method takes the support vector machine as the base classifier, extracts N samples from the original training set with replacement, iterates T times according to this method, and uses the sample set generated each time to train the base classifier model. The classifier encodes real numbers to generate an initial population. In the mutation operation, the difference vector is used to guide the mutation and then produce high-quality individuals. The crossover operation uses the parent individual and the mutant individual to generate crossover individuals, which increases the diversity of individuals. Retention strategies for genetic evolution.