17 results about "Facial appearance" patented technology

A mannequin model video real-person clothing display system based on a diffusion model comprises a data acquisition unit used for acquiring a mannequin model video and a target identity image; and the video generation unit is used for converting the mannequin model video into a real person video with a target identity feature. The video generation unit comprises a feature encoder used for encoding a high-dimensional image and a video into a low-dimensional hidden space; the dynamic posture sensing head encoder is used for fusing the facial appearance features of the target identity image, the facial key point postures of the target identity image and the human body posture sequence features corresponding to the mannequin model video so as to generate posture sensing identity embedding features; the distribution perception adapter is used for carrying out distribution alignment on the identity embedding characteristics and the clothing characteristics; the diffusion video generation network based on the hidden space is used for de-noising reasoning; and the mirror image loss calculation module is used for generating a face region of the video and calculating pixel-level reconstruction loss and perception loss.

The application discloses a face anti-counterfeiting generalization method based on distribution robustness learning and related devices, and acquires a face image to be detected; the face image to be detected is input into a face anti-counterfeiting network model, the face anti-counterfeiting network model outputs a face result, and face anti-counterfeiting generalization is realized; wherein the face anti-counterfeiting network model is obtained by training a distribution enhancement module in a domain and a distribution fusion module between domains. The application enhances the training distribution in various dimensions of main conditions (such as illumination, facial appearance and camera quality) and attack types (such as a new 3D mask made of new materials), so that the enhanced distribution covers part of the test distribution as much as possible, greatly alleviates the problem of distribution deviation, and enables the model to have good generalization capability in unknown attack types and unknown scenes.

A method for detecting an emotional state of a user includes obtaining a first data stream indicative of facial appearance and gaze direction of the user as the user is viewing a scene, determining, based on the first data stream, facial expression feature information indicative of emotional facial expression of the user, obtaining a second data stream indicative of visual content in a field of view of the user, determining, based on the second data stream, visual feature information indicative of visual content in the scene, determining emotional state information based on analyzing the facial expression feature information determined based on the first data stream and the visual feature information determined based on the second data stream, and performing an operation with respect to the emotional state information, wherein the emotional state information is indicative of the emotional state of the user.

Certain aspects and features of this disclosure relate to providing a controllable, dynamic appearance for neural 3D portraits. For example, a method involves projecting a color at points in a digital video portrait based on location, surface normal, and viewing direction for each respective point in a canonical space. The method also involves projecting, using the color, dynamic face normals for the points as changing according to an articulated head pose and facial expression in the digital video portrait. The method further involves disentangling, based on the dynamic face normals, a facial appearance in the digital video portrait into intrinsic components in the canonical space. The method additionally involves storing and / or rendering at least a portion of a head pose as a controllable, neural 3D portrait based on the digital video portrait using the intrinsic components.

The invention discloses a multi-modal analysis traditional Chinese medicine knowledge mining method and system, and the method comprises the steps: obtaining the multi-modal data, such as a tongue image, a face image, a pulse image and an inquiry text, of the same patient, extracting features, and mapping the features to a unified traditional Chinese medicine syndrome semantic space; mining a support, supplement or contradictory relationship among different modal semantic vectors by using an attention mechanism, and constructing a cross-modal evidence collaborative network; generating a comprehensive syndrome representation based on the network fusion information, and completing preliminary differentiation; and finally, generating a final dialectical conclusion and report with credibility evaluation by quantitatively analyzing the consistency degree of the network and the integrity of the evidence chain. The system effectively simulates the thinking process of comprehensive judgment of traditional Chinese medicine doctors, not only improves objectivity and accuracy of syndrome differentiation, but also realizes quantitative evaluation of reliability of diagnosis conclusions and automatic identification of evidence contradictions, and provides an intelligent tool for clinical aid decision making and traditional Chinese medicine experience inheritance.

A method for detecting an emotional state of a user includes obtaining a first data stream indicative of facial appearance and gaze direction of the user as the user is viewing a scene, determining, based on the first data stream, facial expression feature information indicative of emotional facial expression of the user, obtaining a second data stream indicative of visual content in a field of view of the user, determining, based on the second data stream, visual feature information indicative of visual content in the scene, determining emotional state information based on analyzing the facial expression feature information determined based on the first data stream and the visual feature information determined based on the second data stream, and performing an operation with respect to the emotional state information, wherein the emotional state information is indicative of the emotional state of the user

The present application relates to the technical field of image processing, and discloses an image processing method, a device, a storage medium, and a program product. The image processing method comprises: if a portrait to be processed that meets a processing condition (for example, having a feature that affects image aesthetics, such as closed eyes) is present in an image to be processed, acquiring a facial feature of said portrait, and selecting, from an image library, a reference image that has high similarity to the facial feature of said portrait and compensates for said image defect (for example, a reference portrait having the same facial appearance as said portrait and having open eyes); and then performing image restoration on said image on the basis of the reference image to obtain a target image that have overcome said image defect, for example, replacing the part of said image that affects image aesthetics with the corresponding part in the reference image. In this way, the overall aesthetics of the image can be improved, facilitating the improvement of visual experience of a user viewing the target image.

The invention aims to provide a student work auxiliary management device based on multi-modal perception, and the device comprises a multi-modal perception system which carries out the data interaction with a data middle platform, and is used for collecting the basic data of students; the basic data of the student comprises one or more of daily face appearance, wearing, facial micro expression and night abnormal behavior of the student at school. The data processing system performs data interaction with the multi-mode sensing system and is used for acquiring state data of the students in class or night sleeping condition data of the students according to the basic data; according to the state data of the students in class and the night sleeping condition data of the students, abnormal behaviors of the students are judged, and judgment data are obtained; and the application layer performs data interaction with the data processing system and is used for sending reminding information to guardians of the students according to the judgment data so as to perform targeted management on the students.

The present invention discloses an oral administration device, a method for manufacturing the same, and its use. The oral administration device comprises a core assembly and a fixation device, the core assembly comprising a integrally formed molar anastomosis functional region and a drug-loading functional region, the molar anastomosis functional region comprising a first section close to the drug-loading functional region and a second section away from the drug-loading functional region, one end of the first and second sections connected to each other and the other end being an open end, forming a U-shaped structure, the first and second sections used to anastomose with the buccal and lingual sides of a tooth, respectively, the drug-loading functional region comprising a first ring and a second ring arranged coaxially, the fixation device used to enhance the fixation effect of the core assembly to the tooth, and the fixation device is a stator or a fixation layer. When the oral administration device is used, it offers high rigidity, safety, and comfort, and has little impact on facial appearance.

This invention provides a focus recognition method based on an AI-powered education platform. The method includes: capturing input video frames from multiple users; detecting the facial regions of the users; calculating the pixel-averaged image of the set of facial image windows to establish a facial appearance model of the users; generating paths for the users in the multiple input video frames; estimating the detected facial orientation to calculate focus; detecting the number of faces with frontal poses and detecting users who have gazed at the displayed content for a predefined duration; calculating the focus of each user by calculating the time they gaze at the displayed content; associating body language with one of multiple emotion type labels; training a classifier using features extracted from the video frame data; and using the classifier to detect the emotional feedback of the users. This invention proposes a focus recognition method based on an AI-powered education platform, which is better adapted to application scenarios with low-resolution images and combines visual recognition and emotion recognition to help the AI-powered education platform obtain the focus distribution status of users in real time.

The invention discloses a generative skull face restoration method, and relates to the technical field of face restoration, and the method comprises the steps: scanning a tested object, obtaining craniofacial CT data, and carrying out the preprocessing of the CT data; the method comprises the following steps: segmenting outlines of a skull and a wrapper by setting a threshold value, and performing orthogonal projection on three-dimensional data of outline information along a French plane to generate a two-dimensional craniofacial outline image; training a generative adversarial network by using the sample training set, inputting the two-dimensional craniofacial contour image into the trained generative adversarial network, and outputting a facial image of the skull to be restored; the frequency domain sensing generator introduces a semi-wavelet attention mechanism into an encoder and jump connection based on a U-Net network, and the feature characterization capability of a key region is enhanced. According to the invention, by combining the synergistic effect of the half-wavelet attention mechanism in the frequency domain perception generator and the high and low frequency decomposition module, the whole network can simultaneously maintain the accuracy of the overall structure of the face and the definition of the local features in the training process.

The invention discloses an oral administration device as well as a preparation method and application thereof. The oral administration device comprises a core assembly and a fixing device, wherein the core assembly comprises a molar anastomosis functional area and a medicine carrying functional area which are integrally formed; the molar anastomosis functional area comprises a first section close to the medicine carrying functional area and a second section far away from the medicine carrying functional area; one ends of the first section and the second section are mutually connected, and the other ends are open ends to form a U-shaped structure; the first section and the second section are used for being matched with the buccal side and the lingual side of teeth respectively; the medicine carrying functional area comprises a first ring and a second ring which are coaxially arranged; the fixing device is used for enhancing the fixing effect of the core assembly on the teeth; the fixing device is a fixer or a fixing layer. When in use, the oral administration device is high in wearing firmness, high in safety, high in comfort level and small in influence on the appearance of the face.

This invention belongs to the field of computer graphics, and specifically relates to a method and system for capturing facial appearance in outdoor environments based on illumination-reduced priors. The method first constructs a complementary heterogeneous training dataset, which includes a high-fidelity multi-view facial optical property benchmark dataset, a multi-dimensional reflection field dataset for physically realistic rendering, and an HDRI environment map dataset. It then constructs an illumination-reduced prior neural network and trains it using the complementary heterogeneous training dataset, based on latent variable modulation technology of the output dataset and UNet's detail enhancement network, to obtain the trained illumination-reduced prior neural network. Next, it collects outdoor facial appearance data and processes the collected data using the trained illumination-reduced prior neural network to obtain re-illuminated facial assets. The system proposed in this invention can reconstruct high-quality, re-illuminated facial assets from multi-view videos captured by smartphones without human intervention.

From a security perspective, we provide a video processing device, video processing method, and program that can detect fraudulent authentication as early as possible. [Solution] An image processing device according to one embodiment of the present invention comprises: an imaging unit that captures an image; a same-face detection unit that compares a plurality of faces included in the image obtained from the imaging unit and detects that a combination of two or more faces that are presumed to belong to the same person is included; and a same-person determination unit that determines, based on at least one of the location and time of the image capture, that at least one of the faces included in the combination of faces is not an actual appearance of the person.

The application discloses an expression animation redirection method and device, electronic equipment and a computer readable storage medium, wherein the method comprises: obtaining source animation data of each facial skeleton point of a source face model; the topological structure data remains unchanged when the appearance characteristics of the source face model change, and the topological structure data is used for representing the positional relationship between the facial skeleton points; based on the topological structure data and facial appearance characteristics of a target face model of expression redirection, target animation data of each facial skeleton point of the target face model is determined; and position information of each skin vertex of the target face model is determined according to the target animation data. The topological structure data of the facial skeleton points of the source face model is extracted, and the expression animation redirection accuracy is improved.

The application relates to a face image generation method and device, computer equipment and a storage medium. The method comprises the following steps: obtaining a target kinship image of an object having a blood relationship with a target object; determining a target kinship feature corresponding to each structure unit according to the target kinship image; using a preset genetic feature analysis strategy to analyze the target kinship feature corresponding to each structure unit to determine the target genetic feature; and calling a pre-trained face image generation model to process the target genetic feature to generate a face image of the target object. The application can generate a face image of a target object based on a target kinship image of an object having a blood relationship with the target object. The generated face image can highly realistically reflect the current appearance features of a lost or maliciously transferred young object, thereby providing a key and powerful basis for handling such cases.

The invention relates to a face image generation method and apparatus, a computer device and a storage medium. The method comprises the steps of obtaining a target genetic image of an object having a genetic relationship with a target object; determining a target genetic feature corresponding to each structural unit according to the target genetic image; analyzing the target genetic feature corresponding to each structural unit by using a preset genetic feature analysis strategy, and determining the target genetic feature; and calling a pre-trained face image generation model, processing the target genetic features, and generating a face image of the target object. According to the invention, the face image of the target object can be generated based on the target genetic image of the object having the genetic relationship with the target object. The generated face image can highly truly reflect the current facial features of the lost or maliciously transferred young object, and provides a key and powerful basis for handling such cases.