Disclosed implementations include a method, apparatus and computer media for learning an optimal graph in the form of a tree topology defining a sequence that can be used by a learning network for image recognition. Image data representing the image of an object is received and N landmarks are detected on the image using a deep regression algorithm, wherein N is an integer. A weighted, fully connected, graph is constructed from the landmarks by assigning initial weights for the landmarks randomly. An optimized tree structure is determined based on the initial weights. A sequence is generated by traversing nodes of the tree structure and a series of embeddings representing the object image are generated based on the sequence. The embeddings can be processed by a neural network to generate an image recognition signal based on the embeddings.

An information processing apparatus comprising a situation acquiring unit configured to acquire a situation of one movable object, and a storing execution unit configured to store a face image of an occupant on the one movable object, captured by an image-capturing unit mounted on the one movable object when the one movable object is in a predetermined situation, as a face image of the occupant with neutral facial expression is provided.

A system and method of animating an image of an object by extracting a plurality of three dimensional (3D) features from a first image, depicting a puppet object, and from the second image of a driver object, obtaining, a value of a first 3D identity-invariant feature of the puppet object, and a value of a second 3D identity-invariant feature, from the second image, calculating and applying, a mixing function, on the plurality of 3D features, generating a rendered image, based on the plurality of 3D features, the first 3D identity-invariant feature and the second 3D identity-invariant feature, and generating an output image based on the rendered image, the plurality of 3D features, the first image and the second image, wherein the output image depicting a target object comprising at least one 3D identity-invariant feature of the driver object.

A system which, with data provided by one or more sensors, detects a user's alteration of the geometries of parts of his face, head, neck, and/or shoulders. It determines the extent of each alteration and normalizes it with respect to the maximum possible range of each alteration so as to assign to each part-specific alteration a numeric score indicative of its extent. The normalized part-specific scores are combined so as to produce a composite numeric code representative of the complete set of simultaneously-executed geometric alterations. Each composite code is translated, or interpreted, relative to an appropriate context defined by an embodiment, an application executing on an embodiment, or by the user. For example, each composite code might be interpreted as, or assigned to, a specific alphanumeric letter, a color, a musical note, etc. Through the use of this system, a user may communicate data and/or commands to a computerized device, while retaining full use of his hands and his voice for other tasks, and while being free to focus his visual attention on something other than the system.

The innovation disclosed and claimed herein, in aspects thereof, comprises systems and methods of identifying AUs and emotion categories in images. The systems and methods utilized a set of images that include facial images of people. The systems and methods analyze the facial images to determine AUs and facial color due to facial blood flow variations that are indicative of an emotion category. In aspects, the analysis can include Gabor transforms to determine the AUs, AU intensities and emotion categories. In other aspects, the systems and method can include color variance analysis to determine the AUs, AU intensities and emotion categories. In further aspects, the analysis can include convolutional neural networks that are trained to determine the AUs, emotion categories and their intensities.

An image recognition method is provided, including: obtaining a target video including a target object; extracting a target video frame image from the target video; generating a key point video frame sequence comprised of a plurality of key point video frames according to key point information of the object and a plurality of video frames in the target video; extracting dynamic timing feature information of the key point video frame sequence; extracting static structural feature information of the target video frame image; and recognizing an attribute type corresponding to the target object in the target video according to the dynamic timing feature information of the key point video frame sequence and the static structural feature information of the target video frame image.

Provided are systems and a method for photorealistic real-time face reenactment. An example method includes receiving a target video including a target face and a scenario including a series of source facial expressions, determining, based on the target face, one or more target facial expressions, and synthesizing, using the parametric face model, an output face. The output face includes the target face. The one or more target facial expressions are modified to imitate the source facial expressions. The method further includes generating, based on a deep neural network, a mouth region and an eyes region, and combining the output face, the mouth region, and the eyes region to generate a frame of an output video.

The invention provides a method and device for protecting a child inside a vehicle, a computer device, a computer-readable storage medium, and a vehicle, which are applied to the technical field of automobiles. The method for protecting a child inside a vehicle includes: receiving image data inside a vehicle; performing facial detection on the image data to detect passengers in the vehicle; further classifying the detected passengers in the vehicle into children and adults according to the facial detection; and performing a corresponding protection operation when a result of the classification indicates that there is a child alone in the vehicle.

A method and for child state analysis, a vehicle, an electronic device, and a storage medium are provided. The method includes: performing face feature extraction on at least one image frame in an obtained video stream; classifying whether a person in the image is a child and at least one state of the person according to face features to obtain a first classification result of whether the person in the image is a child, and a second classification result of the at least one state of the person; outputting the first classification result and the second classification result; and/or outputting prompt information according to the first classification result and the second classification result.

An artificial intelligence (AI)-assisted evaluation method for aesthetic medicine and an evaluation system are provided. An AI aesthetic medicine identification and analysis module is used. An AI facial expression evaluation module provides a real-time facial expression evaluation result of a subject. The real-time facial expression evaluation result is inputted into the AI aesthetic medicine identification and analysis module. The AI aesthetic medicine identification and analysis module optionally cooperates with at least one of a medical knowledge rule module and an aesthetic medicine auxiliary evaluation result historical database to perform an AI aesthetic medicine identification and analysis process. Then, the AI aesthetic medicine identification and analysis module generates and outputs a real-time aesthetic medicine auxiliary evaluation result. According to the real-time aesthetic medicine auxiliary evaluation result, an aesthetic medicine behavior is carried out. Consequently, the personalized aesthetic therapeutic effect can be achieved.