Abstract: An identity verification method performed at a terminal includes playing in an audio form action guide information including mouth shape guide information selected from a preset action guide information library at a speed corresponding to the action guide information, and collecting a corresponding set of action images within a preset time window; performing matching detection on the collected set of action images and the action guide information, to obtain a living body detection result indicating whether a living body exists in the collected set of action images; according to the living body detection result that indicates that a living body exists in the collected set of action images: collecting user identity information and performing verification according to the collected user identity information, to obtain a user identity information verification result; and determining the identity verification result according to the user identity information verification result.

Abstract: Face model matrix training method, apparatus, and storage medium are provided. The method includes: obtaining a face image library, the face image library including k groups of face images, and each group of face images including at least one face image of at least one person, k>2, and k being an integer; separately parsing each group of the k groups of face images, and calculating a first matrix and a second matrix according to parsing results, the first matrix being an intra-group covariance matrix of facial features of each group of face images, and the second matrix being an inter-group covariance matrix of facial features of the k groups of face images; and training face model matrices according to the first matrix and the second matrix.

Abstract: A face liveness detection method includes outputting a prompt to complete one or more specified actions in sequence within a specified time period, obtaining a face video, detecting a reference face image frame in the face video using a face detection method, locating a facial keypoint in the reference face image frame, tracking the facial keypoint in one or more subsequent face image frames, determining a state parameter of one of the one or more specified actions using a continuity analysis method according to the facial keypoint, and determining whether the one of the one or more specified actions is completed according to a continuity of the state parameter.

Abstract: An identity verification method performed at a terminal includes: displaying and/or playing in an audio form action guide information selected from a preset action guide information library, and collecting a corresponding set of action images within a preset time window; performing matching detection on the collected set of action images and the action guide information, to obtain a living body detection result indicating whether a living body exists in the collected set of action images; according to the living body detection result that indicates that a living body exists in the collected set of action images: collecting user identity information and performing verification according to the collected user identity information, to obtain a user identity information verification result; and determining the identity verification result according to the user identity information verification result.

Abstract: Disclosed are an evaluation method and an evaluation device for a facial key point positioning result. In some embodiments, the evaluation method includes: acquiring a facial image and one or more positioning result coordinates of a key point of the facial image; performing a normalization process on the positioning result coordinate and an average facial model to obtain a normalized facial image; and extracting a facial feature value of the normalized facial image and calculating an evaluation result based on the facial feature value and a weight vector.

Abstract: Face model matrix training method, apparatus, and storage medium are provided. The method includes: obtaining a face image library, the face image library including k groups of face images, and each group of face images including at least one face image of at least one person, k>2, and k being an integer; separately parsing each group of the k groups of face images, and calculating a first matrix and a second matrix according to parsing results, the first matrix being an intra-group covariance matrix of facial features of each group of face images, and the second matrix being an inter-group covariance matrix of facial features of the k groups of face images; and training face model matrices according to the first matrix and the second matrix.

Abstract: A statistical parameter modeling method is performed by a server. After obtaining model training data, the model training data including a text feature sequence and a corresponding original speech sample sequence, the server inputs an original vector matrix formed by matching a text feature sample point in the text feature sample sequence with a speech sample point in the original speech sample sequence into a statistical parameter model for training and then performs non-linear mapping calculation on the original vector matrix in a hidden layer, to output a corresponding prediction speech sample point. The server then obtains a model parameter of the statistical parameter model according to the prediction speech sample point and a corresponding original speech sample point by using a smallest difference principle, to obtain a corresponding target statistical parameter model.

Abstract: A facial tracking method is provided. The method includes: obtaining, from a video stream, an image that currently needs to be processed as a current image frame; and obtaining coordinates of facial key points in a previous image frame and a confidence level corresponding to the previous image frame. The method also includes calculating coordinates of facial key points in the current image frame according to the coordinates of the facial key points in the previous image frame when the confidence level is higher than a preset threshold; and performing multi-face recognition on the current image frame according to the coordinates of the facial key points in the current image frame. The method also includes calculating a confidence level of the coordinates of the facial key points in the current image frame, and returning to process a next frame until recognition on all image frames is completed.

Abstract: Disclosed are an evaluation method and an evaluation device for a facial key point positioning result. In some embodiments, the evaluation method includes: acquiring a facial image and one or more positioning result coordinates of a key point of the facial image; performing a normalization process on the positioning result coordinate and an average facial model to obtain a normalized facial image; and extracting a facial feature value of the normalized facial image and calculating an evaluation result based on the facial feature value and a weight vector.

Abstract: An image classification method is provided. The method includes: inputting a to-be-classified image into a plurality of neural network models; obtaining data output by multiple non-input layers specified by each neural network model to generate a plurality of image features corresponding to the plurality of neural network models; respectively inputting the plurality of corresponding image features into linear classifiers, each of the linear classifiers being trained by one of the plurality of neural network models for determining whether an image belongs to a preset class; obtaining, using each neural network model, a corresponding probability that the to-be-classified image comprises an object image of the preset class; and determining, according to each obtained probability, whether the to-be-classified image includes the object image of the preset class.

Abstract: A method and an apparatus for training a voiceprint recognition system are provided. The method includes obtaining a voice training data set comprising voice segments of users; determining identity vectors of all the voice segments; identifying identity vectors of voice segments of a same user in the determined identity vectors; placing the recognized identity vectors of the same user in the users into one of user categories; and determining an identity vector in the user category as a first identity vector. The method further includes normalizing the first identity vector by using a normalization matrix, a first value being a sum of similarity degrees between the first identity vector in the corresponding category and other identity vectors in the corresponding category; training the normalization matrix, and outputting a training value of the normalization matrix when the normalization matrix maximizes a sum of first values of all the user categories.

Abstract: An identity vector generation method is provided. The method includes obtaining to-be-processed speech data. Corresponding acoustic features are extracted from the to-be-processed speech data. A posterior probability that each of the acoustic features belongs to each Gaussian distribution component in a speaker background model is calculated to obtain a statistic. The statistic is mapped to a statistic space to obtain a reference statistic, the statistic space built according to a statistic corresponding to a speech sample exceeding a threshold speech duration. A corrected statistic is determined according to the calculated statistic and the reference statistic; and an identity vector is generated according to the corrected statistic.

Abstract: A sign-in method and server based on facial recognition are provided. The method includes: receiving a face image of a sign-in user from a sign-in terminal. According to the face image of the sign-in user, whether a target registration user matching the sign-in user exists in a pre-stored registration set is detected. The registration set includes a face image of at least one registration user. Further, the target registration user is confirmed as signed in successfully if the target registration user exists in the registration set.

Abstract: A voice data processing method and apparatus are provided. The method includes obtaining an I-Vector vector of each of voice samples, and determining a target seed sample in the voice samples. A first cosine distance is calculated between an I-Vector vector of the target seed sample and an I-Vector vector of a target remaining voice sample, where the target remaining voice sample is a voice sample other than the target seed sample in the voice samples. A target voice sample is filtered from the voice samples or the target remaining voice sample according to the first cosine distance, to obtain a target voice sample whose first cosine distance is greater than a first threshold.

Abstract: The embodiments of the present disclosure disclose a method for constructing a speech decoding network in digital speech recognition. The method comprises acquiring training data obtained by digital speech recording, the training data comprising a plurality of speech segments, and each speech segment comprising a plurality of digital speeches; performing acoustic feature extraction on the training data to obtain a feature sequence corresponding to each speech segment; performing progressive training starting from a mono-phoneme acoustic model to obtain an acoustic model; acquiring a language model, and constructing a speech decoding network by the language model and the acoustic model obtained by training.

Abstract: The present disclosure pertains to the field of image processing technologies and discloses a face key point positioning method and a terminal. The method includes: obtaining a face image; recognizing a face frame in the face image; determining positions of n key points of a target face in the face frame according to the face frame and a first positioning algorithm; performing screening to select, from candidate faces, a similar face whose positions of corresponding key points match the positions of the n key points of the target face; and determining positions of m key points of the similar face selected through screening according to a second positioning algorithm, m being a positive integer. In this way, the problem that positions of key points obtained by a terminal have relatively great deviations in the related technologies is resolved, thereby achieving an effect of improving accuracy of positioned positions of the key points.

Abstract: A 3D human face reconstruction method and apparatus, and a server are provided. In some embodiments, the method includes determining feature points on an acquired 2D human face image; determining posture parameters of a human face according to the feature points, and adjusting a posture of a universal 3D human face model according to the posture parameters; determining points on the universal 3D human face model corresponding to the feature points, and adjusting the corresponding points in a sheltered status to obtain a preliminary 3D human face model; and performing deformation adjusting on the preliminary 3D human face model, and performing texture mapping on the deformed 3D human face model to obtain a final 3D human face.

Abstract: A melody information processing method is described. A piece of Musical Instrument Digital Interface (MIDI) data corresponding to a song is received, a song identifier of the song is obtained, first melody information is generated according to the MIDI data, and the first melody information is stored in association with the song identifier in a melody database. Moreover, a user unaccompanied-singing audio data set that is uploaded from a user terminal is received, second melody information corresponding to the song identifier is extracted according to the user unaccompanied-singing audio data set, and the second melody information is stored in association with the song identifier in the melody database.

Abstract: Processing circuitry of an information processing apparatus obtains a set of identity vectors that are calculated according to voice samples from speakers. The identity vectors are classified into speaker classes respectively corresponding to the speakers. The processing circuitry selects, from the identity vectors, first subsets of interclass neighboring identity vectors respectively corresponding to the identity vectors and second subsets of intraclass neighboring identity vectors respectively corresponding to the identity vectors. The processing circuitry determines an interclass difference based on the first subsets of interclass neighboring identity vectors and the corresponding identity vectors; and determines an intraclass difference based on the second subsets of intraclass neighboring identify vectors and the corresponding identity vectors.

Abstract: Aspects of the disclosure provide a method for adding a target contact to a user's friend list in a social network. A target image of a human body part of the target contact can be received from a user terminal. A target biological feature can be extracted from the target image. Whether the target biological feature matches a reference biological feature of a plurality of prestored reference biological features can be determined. A social account associated with the determined reference biological feature that matches the target biological feature may be determined, and added to the user's friend list.