433 results about "Face contour" patented technology

The invention discloses a human face super-resolution reconstruction method based on a generative adversarial network and sub-pixel convolution, and the method comprises the steps: A, carrying out the preprocessing through a normally used public human face data set, and making a low-resolution human face image and a corresponding high-resolution human face image training set; B, constructing the generative adversarial network for training, adding a sub-pixel convolution to the generative adversarial network to achieve the generation of a super-resolution image and introduce a weighted type loss function comprising feature loss; C, sequentially inputting a training set obtained at step A into a generative adversarial network model for modeling training, adjusting the parameters, and achieving the convergence; D, carrying out the preprocessing of a to-be-processed low-resolution human face image, inputting the image into the generative adversarial network model, and obtaining a high-resolution image after super-resolution reconstruction. The method can achieve the generation of a corresponding high-resolution image which is clearer in human face contour, is more specific in detail and is invariable in features. The method improves the human face recognition accuracy, and is better in human face super-resolution reconstruction effect.

A system and method are capable of performing make-up virtual images obtained with portable devices with a digital camera. The method automatically detects points of interest (eyes, mouth, eyebrow, face contour) of the user's face image allowing the user to virtually apply make-up using fingers on a touch screen. Another algorithm was created to avoid “blurring” of make-up in the application from the user's fingers. The system allows the creation of virtual make-up faces with high accuracy. To carry out the virtual make-up, methods were created for automatically detecting points of interest in the facial image and enforcing transparency in the image to simulate the make-up and restrict the make-up application area within the region where the points of interest are found. The system allows testing cosmetics application in facial images to aid choosing colors of cosmetics before actually applying them.

A protective goggle assembly has a goggle frame having a rigid goggle body formed on a front side. The goggle body has recesses formed on two sides of a lower end so that a resilient face engaging portion is integrally formed with a rear side of the goggle frame. The resilient face engaging portion includes edges on two sides of a lower end, which edges are formed with a curved shape based on a human face contour. When the goggle frame and the lens are assembled with a strap buckled to the two sides for use, and the resilient face engaging portion on the rear side of the goggle frame is in contact with a user's face and cheekbones, the recesses formed on the two sides of the lower end of the rigid goggle body provide the resilient face engaging portion with a larger space for flexible adjustment.

The invention relates to a triangular gridding based human image face-lift processing method and belongs to the technical field of computer application. By means of triangulation and feature point extraction technology and related geometric method to process the image, the triangular gridding based human image face-lift processing method includes steps of feature point extraction of human faces, triangular gridding, human face outline curve estimation, feature point adjustment, new triangular grid establishing, new image generation and the like. Considering beauty seeking of people and blank of related technology, the triangular gridding based human image face-lift processing method is an automatic digital image face-lift beautifying method, and can realize face-lift beautifying effect simply and quickly by loading in original source image to be subjected to face lift, so that beauty appreciation of people is sufficiently satisfied without undertaking high cost and risk.

The invention provides a method for evaluating beautifulness of human faces based on video. The method is achieved by following steps: (1) capturing the face information by video; (2) effectively detecting the face according to the captured information by a cascade classifier based on haar-like characteristics; locating the eyes and the lips on the basis of the face detection, including detecting eyes by adopting a Sobel edge detection operator and a region growing method and determining the boundary of the lips according to the projection of red pixel dots in the face image; (4) extracting the face contour by adopting Snake model; and (5) comparing the extracted information with a face beautifulness standard library to determine the face beautifulness evaluation result. The method can achieve the real-time face beautifulness evaluation by capturing the face video information by a computer, and training to obtain the beautifulness classification standard according to a large amount of face samples.

The invention provides a face motion tracking method, a face motion tracking system and a method for enhancing reality. The face motion tracking method comprises the following steps of: 1) extracting face characteristic points and face contour characteristics of a video image; 2) determining the number of the face characteristic points; and 3) under the condition that the number of the face characteristic points is greater than a preset threshold value, performing motion tracking on faces in the video images by using the face characteristic points and a three-dimensional model, and under the condition that the number of the face characteristic points is smaller than or equal to the preset threshold value, performing the motion tracking by using the face contour characteristics. By the method, head motions can be accurately tracked to prevent determined face characteristic points from being lost; moreover, during real-time tracking, only a few face motion data is transmitted by a communication network, so the method is suitable for various mobile phone platforms, has low requirements on hardware conditions, and can be applied to video call interaction to improve the interest and the interaction of a call.

The invention relates to an image processing method, in particular to a detecting and classifying method of facial-form contours. The invention aims at solving the problems in the prior art that the accurate facial-form contour can not detected and the techniques can not be really applied to the facial-form classification. The invention provides the technical proposal that: the step 1, the human face is positioned; the step 2, the facial contour is detected: an egg-shaped face is set into a vector closed and symmetrical curve Lm that consists of one or a plurality of points, and then, the forehead height, the chin height, the abscissa position that keeps to the sides of the left cheek and the right cheek extremely and the gill shape are modified; the step 3, the facial forms are classified. The invention can provide the effective and automatic image processing measures for a plurality of industries such as image design, facial make-up, hairdressing, jewelry, face-painting, entertainment, and the like, and provides the automated and objective technical support for a plurality of applied industries.

A biaxial razor is disclosed. The biaxial razor includes a shaving head, a handle, and a pivot assembly arranged between the shaving head and the handle. The pivot assembly is composed of a first axial driving group and a second axial driving group. The first and second axial driving groups enable the shaving head to make both a rotating displacement in a front-rear direction and a swing displacement in a left-right direction, so as to perfectly fit with facial skins of users whose face contours are different, without taking lots of arm motions in operations when shaving, thereby to enhance the convenience in operations and improve the efficiency in shaving.

The invention discloses an image processing method and device. According to the embodiment, the method comprises the steps: detecting an input human face area in a portrait, and obtaining human face contour information corresponding to the human face area, the human face contour information containing human face contour key points in the human face area; obtaining the human face contour dimension corresponding to the human face contour information on the basis of the human face contour information; when the human face contour dimension is identified to be within a preset face dimension range, obtaining a corresponding cartoon face material according to the human face contour information; and adjusting the cartoon face material on the basis of the human face contour key points, and obtaining cartoon face information after deformation so that a corresponding cartoon head portrait is synthesized according to the deformed cartoon face information. According to the embodiment, the image processing method and device have the beneficial effects that a cartoon face corresponding to a human face in a portrait picture can be obtained in a self-adaptive manner, and similarity and authenticity between a cartoon face and a real human face are improved.

A feature based face recognition method and apparatus. Global amplitude and direction is calculated for regions defined by a sliding window over an input gray scale image. These values are low pass filtered and a search is carried out for candidates for an eye. Once an eye candidate is identified by comparing the amplitude with a threshold, the other eye is presumed to be located a fixed distance away at the angle of the eye. The mouth is then determined to be in a predetermined location perpendicular to the angle of the eyes. A face contour is estimated by defining two ellipses between which face contours having angles similar to the ellipse are found. Multiple detections are possible and those detections are merged by averaging. The image is resized and the process repeats until the image has been reduced to its minimum size.

The invention discloses an image block weighted convolutional neural network-based face recognition method. The method includes the following steps that: a sample database is constructed, edge detection is performed on a sample picture, and a face contour picture is intercepted; five portions in the face contour image are positioned, segmentation is performed according to positioning, so that local pictures can be obtained, and the gray scale variance mean of the local pictures is calculated; the face contour picture and six local pictures, which belong to the same sample picture, are together put into the training of a convolutional neural network; and segmented pictures of a pictures to be recognized are put into the trained convolutional neural network, so that a recognition result can be obtained. According to the image block weighted convolutional neural network-based face recognition method of the invention, both local characteristics and global characteristics are considered, so that a system can have a better recognition effect. According to the image block weighted convolutional neural network-based face recognition method of the invention, the image block weighted convolutional neural network is adopted, and thus, compared with a traditional face recognition method, the image block weighted convolutional neural network-based face recognition method of the present invention can improve the recognition rate of face recognition.

Disclosed are systems and methods for generating large data sets for training deep learning networks for 3D measurements extraction from images taken using a mobile device camera. The method includes the steps of receiving at least one 3D model; generating a first type of augmentation data, such as but not limited to skin color, face contour, hair style, virtual clothing, and / or lighting conditions; augmenting the 3D model with the first type of augmentation data; generating at least one image from the augmented 3D model; receiving a second type of augmentation data, such as a plurality of background images representing a variety of backgrounds; augmenting the at least one image with the second type of augmentation data to generate a plurality of augmented images; extracting spatial features from the 3D model; and providing the plurality of augmented images and the spatial features to train a deep learning network for 3D measurement determination.

The invention provides an image processing method and device, a readable storage medium and computer equipment. The image processing method comprises the steps that size information of a forehead region in a face region of a to-be-processed image is acquired; when the size information of the forehead region is greater than a preset scale value of size information of the face region, a face contourof the face region is acquired; a face type conforming to the face contour is acquired according to a preset face sample library; and a hairstyle capable of fusing with a hair region of the to-be-processed image is generated according to the face type, and a synthetic image is obtained. Through the image processing method, a wig capable of fusing with the hair region can be configured for the hair region of the to-be-processed image, so that the image displays a better effect.

The invention discloses a multi-neural network cascaded face recognition method based on human face key points. The multi-neural network cascaded face recognition method comprises the steps of: detecting a face image by adopting an MTCNN algorithm, and then utilizing affine transformation to rotate, translate and zoom a human face for subsequent processing; and utilizing a convolutional neural network for detecting human face contour key points and human face internal key points respectively, and performing feature dimensionality reduction by adopting a PCA algorithm. When performing dimensionality reduction, methods based on a category pattern can be adopted according to different categories, the problems that the traditional PCA algorithm cannot effectively utilize category information between categories and the robustness is poor in the presence of illumination and facial expression changes can be overcame, and finally, the BP neural network is used for recognition. Since advanced or improved methods are used in many links, good effect can be obtained when face recognition is performed.

The invention relates to a method for automatically converting a single face front photo into a three-dimensional (3D) face model. The method is characterized by comprising the steps of a, inputting a front people picture; b, extracting and partitioning characteristic points; c, establishing a standard 3D face model; d, dividing vertexes of the standard face model into groups, and separating out point groups of the five sense organs, wherein the point groups of the five sense organs can be selectively formed by vertex groups of eyebrows, eyes, a nose, a mouth and a face contour; e, performing overall zooming on the standard 3D face model by using a vertex position adjusting module of the standard 3D face model; f, adjusting the vertex position adjusting module of the standard 3D face model; g, finishing the reconstruction of the entire standard 3D face model. The method provided by the invention can be used for automatically converting one face front photo into a 3D face and restoring a real face to the maximum extent in a way of according with face biological characteristics so as to facilitate subsequent animation processing.

The invention discloses a single photograph-based fast face modeling method. The method includes the following steps that: the three-dimensional model base of various standard faces is established; facial feature point recognition is performed on an inputted face photograph through a face recognition technology; the face photograph is zoomed, translated and rotated on the basis of facial feature points, so that the face of the photograph is preliminary overlapped with face grids under a uv plane; a feature face is completely overlapped with a uv face; a facial feature region is extracted, and Poisson fusion is performed on the facial feature region and an original texture map, so that a new map can be formed; whitening and dermabrasion are performed on the new map; a weight is solved according to face contour parameters and face contour feature points, a new three-dimensional face model is constructed; and texture mapping is performed on the new face map and the new three-dimensional face model, so that a complete three-dimensional face model can be obtained. With the method of the invention adopted, by means of a single photograph, a three-dimensional face model close to the photograph can be generated fast.

A personalized protective mask comprises a malleable wire frame, a filtration bag, and a tether. The filtration bag is positioned over the wire frame and the tether is attached at each end to an ear section of the wire frame. The mask is placed over the face of the individual who then bends the wire frame to force it to conform to the individual's own face contour. The tether pulls tight the filtration bag to the individual's face for creating a mask which fully covers the user's nose and mouth. The protective mask is comfortable to wear, whether in-use on the face or in a stand-by position around the user's neck.

A system and method are capable of performing make-up virtual images obtained with portable devices with a digital camera. The method automatically detects points of interest (eyes, mouth, eyebrow, face contour) of the user's face image allowing the user to virtually apply make-up using fingers on a touch screen. Another algorithm was created to avoid “blurring” of make-up in the application from the user's fingers. The system allows the creation of virtual make-up faces with high accuracy. To carry out the virtual make-up, methods were created for automatically detecting points of interest in the facial image and enforcing transparency in the image to simulate the make-up and restrict the make-up application area within the region where the points of interest are found. The system allows testing cosmetics application in facial images to aid choosing colors of cosmetics before actually applying them.

Provided is a 3D precision shaping method of a conformal cooling structure of an injection mold. Based on heat transfer calculation, an efficient heat exchange conformal cooling flow channel three-dimensional model is established for the cavity face contour of the injection mold, layered structural division is performed on the conformal cooling flow channel three-dimensional model, corresponding structures in different layers are machined through a machining method, a laser method or a chemical etching method, and the structures in different layers are stacked and positioned according to corresponding sequence. The layered structures are welded through a diffusion welding method achieved by adding an intermediate layer, and 3D precision shaping of a conformal cooling flow channel is achieved; finally, a mold cavity is machined in a diffusion welding solid surface through a machining method, and 3D precision shaping of the conformal cooling structure of the injection mold is completed. By means of the method, the design space can be greatly expanded, and conformal cooling structures in any complicated shape can be manufactured. High strength and high precision of the conformal cooling structure is guaranteed, production cost is low, production efficiency is high, and the requirement for batch industrial production can be met.

Disclosed is a three-dimensional face recognition method based on face contour lines of semi-rigid areas. The method comprises steps of (1) preprocessing three-dimensional face models, including face area segmentation, smoothing and pose normalization, and placing all faces under a pose coordinate system; (2) extracting multiple vertical contour lines of the faces from the semi-rigid areas of the three-dimensional models to represent the facial curved surfaces of the faces; (3) utilizing an elastic curve matching algorithm to compute the geodesic distances among the corresponding contour lines of different three-dimensional face models in a pre-shape space, taking the geodesic distances as similarity measures, weight-fusing the similarity vectors of all the contour lines to obtain a total similarity for recognition. The three-dimensional face recognition method based on the face contour lines of the semi-rigid areas has high recognition performance, achieves high robustness on change in expression and increases the face matching speed.

The invention discloses a local region matching-based face search method. The method includes the following steps that: 1) faces of each image in a face image set A are aligned with a face of a standard format, and areas of various organs are divided; 2) bottom-level feature vectors of each organ are extracted from and are clustered; 3) any two classifications are selected from clustering results of each organ and are adopted as positive and negative samples, and a support vector machine classifier is trained; training is performed in a paired combination manner, such that a classifier set of the organs can be obtained, and the results of discrimination of the bottom-level feature vectors which is performed by each classifier in the classifier set are united so as to form new feature vectors, namely, middle-level feature vectors of the organs; 4) the ratio of the distance of each key point on each face contour to left and right eyes to the distance between the two eyes is calculated and is adopted as the middle-level feature vector of the corresponding face contour; the above middle-level feature vectors are combined such that Vr can be obtained; and 5) a middle-level feature vector Vq is generated for a face image q to be searched; and the Vq is matched with the Vr in the A, and query results are returned. With the local region matching-based face search method of the invention adopted, a search effect of similar faces can be improved.

The embodiment of the invention discloses a face detection method and device. The method comprises the following steps: carrying out edge detection processing on a preprocessed image obtained by preprocessing; obtaining an edge detection area and the length value, the width value and the length-width ratio of the edge detection area; according to the length value, the width value and the length-width ratio, obtaining a round area to be detected; carrying out expanded processing on the complexion area to the preprocessed image within the coverage range of the round area to be detected, and obtaining a first face contour; carrying out light compensation processing and complexion detection on the first light compensation image within the coverage range of the round area to be detected to obtain a second face contour; and combining a first face contour with a second face contour to obtain the face image. The device comprises a contour detection module, a complexion area expansion module, a light compensation and complexion detection module and a combination module. The embodiment of the invention avoids the phenomenon that face data is lost because of shadow caused by shading in the complexion detection process, and the accuracy rate and the detection speed of face detection can be improved.

The invention discloses a face feature point positioning method and device, so as to solve the problems of inaccurate positioning and low speed of the face feature point positioning method. The methodcomprises steps: a to-be-detected face image is acquired; through a first convolution layer in an ith convolution module of a deep convolution neural network model, a face feature image processed andobtained by an (i-1)th convolution module is fused based on multiple groups of weights, and a first feature image is obtained; through a second convolution layer in the ith convolution module, the first feature image is fused again based on multiple groups of weights, and a second feature image is obtained; through a third convolution layer in the ith convolution module, face contour features andfive sense organ features in the first feature image are extracted to obtain a third feature image; through a fusion layer in the ith convolution module, the second feature image and the third feature image are fused; and a face feature image processed and obtained by a final convolution module of the deep convolution neural network model is used as the face feature image of the to-be-detected face image.

The present invention relates to a face image preferential system. The system comprises a rotation calculation module for obtaining a rotation degree of a face in each face image, an image quality calculation module for obtaining the definition and face face contour quantity of each face image, a face state judgment module for judging whether two eyes in each face image are opened or not and judging whether a mouth in each face image is shut up or not, and a score calculation module for obtaining value sequencing of each face image. The present invention further relates to a face image preferential processing method. The system and the method of the present invention improves the identification accuracy of face images, can be used for excluding and selecting an optimal face figure, and have relatively good generalization.

The method of covering face of news interviewee via quick face detection includes the following steps: detecting the position and size of the news interviewee; verifying and compensating with complexion information; tracking the face of news interviewee with face complexion histogram; and extracting face contour and blurring processing by means of the facial position, region and complexion information. The automatic face detecting and covering technology can reach a face detecting rate up to 88 % in case of using no complexion information and cover almost the whole face while combining with the complexion information. Therefore, the present invention can reduce work strength and raise work efficiency.

The invention discloses a human-face-based real-time depth of field method. The method comprises the steps that real-time images are obtained by obtaining real-time viewing camera data, human face detection is carried out on the real-time images to obtain human face areas and human face key points, then, affine transformation is adopted for obtaining transformed human face contour images according to preset human face contour images, preset human face key points corresponding to the human face contour images, and the obtained human face areas and human face key points, finally, the transformed human face contour images are used as the masking of the human face areas of the real-time images for conducting transparency calculation on the real-time images and blurred images to obtain human face depth of field images, and the human face depth of field images are used as display images to be previewed on a screen in real time. No hardware cost is needed, and no manual intervention operation is needed, so that the automatic depth of field effect processing during self-photographing is achieved, and the self-photographing effect is better and more natural.

The invention discloses a public transport travel control method and device, a computer device and a storage medium. The method comprises: obtaining face images of one or more passengers entering a public transportation jurisdiction area, and enabling the face images to comprise face contour characteristics; matching a face image in a preset database, wherein the preset database comprises the collected face image and a set of identity information and account information mapped with the face image; and detecting whether the face image satisfies a preset condition, and when the face image satisfies the preset condition, allowing passage and paying the traffic fee from the account information mapped by the face image. The whole process is simple and convenient, multiple persons can be read atthe same time, the recognition speed and the passing speed of the passengers are increased, the riding time is saved, automatic payment is achieved, and the passengers can ride more conveniently. Byobtaining the identity information of the passengers, the passengers can be monitored conveniently, lawbreakers are prevented from being mixed into public transportation places to be disordered, and the public transportation safety order supervision strength is improved.

The invention discloses an animal face key point marking method. The method includes the steps of detecting an animal face of an input sample image to obtain an animal face contour and the positions of animal face key points, processing the input sample image according to the animal face contour and the positions of the animal face key points, sending the processed sample image into a pre-trainedkey point prediction network to obtain a first key point prediction result, adjusting the first key point prediction result to obtain marked points of the animal face key points, and conducting quality detecting on the marked points according to a pre-defined standard image to obtain the quality score of the marked points. By means of the technical scheme, the animal face key points can be precisely marked, the marked points can be automatically checked, and therefore the marking precision is improved and the work efficiency is greatly improved.

The invention provides a human face video image based heart rate detection system which comprises a video recording module, a face recognition module, an RGB separation module, an independent component analysis module and a physiological parameter quantization module. The video recording module, the face recognition module, the RGB separation module, the independent component analysis module and the physiological parameter quantization module are connected successively. The heart rate detection methods include steps of video recording, face recognition, RGB separation, independent component analysis and decomposition, physiological parameters quantification. The system and the method have the advantages of high calculation efficiency, effectively overcoming the influence of human face contour, light and other environmental factors. The system can also remind real-time cardiovascular changes, solve the problem of inaccurate measuring results due to the discomfort of a patient to contact detection, and has a large application prospect in remote medical care.