Abstract: A processor-implemented method with neural network training includes: determining first backbone feature data corresponding to each input data by applying, to a first neural network model, two or more sets of the input data of the same scene, respectively; determining second backbone feature data corresponding to each input data by applying, to a second neural network model, the two or more sets of the input data, respectively; determining projection-based first embedded data and dropout-based first view data from the first backbone feature data; and determining projection-based second embedded data and dropout-based second view data from the second backbone feature data; and training either one or both of the first neural network model and the second neural network model based on a loss determined based on a combination of any two or more of the first embedded data, the first view data, the second embedded data, the second view data, and an embedded data clustering result.

Abstract: A processor-implemented method includes: initializing a neural network model with arbitrary values using a random seed; training the neural network model based on the arbitrary values; determining a number of coats and respective densities of the coats; learning respective scores of parameters of the neural network model based on the number of coats and the respective densities of the coats; determining mask information for determining the parameters of the neural network model to be comprised in each of the coats based on the scores; and generating a bitstream based on the number of coats, the respective densities of the coats, the mask information, and the random seed.

Abstract: A training method of a neural network, and a recognition method and apparatus using the neural network are disclosed. The recognition method using the neural network includes obtaining a feature vector generated from a hidden layer of the neural network, in response to data being entered to an input layer of the neural network, and determining a reliability of a recognition result for the data using the feature vector and clusters.

Abstract: A processor-implemented method with neural network training includes: determining first backbone feature data corresponding to each input data by applying, to a first neural network model, two or more sets of the input data of the same scene, respectively; determining second backbone feature data corresponding to each input data by applying, to a second neural network model, the two or more sets of the input data, respectively; determining projection-based first embedded data and dropout-based first view data from the first backbone feature data; and determining projection-based second embedded data and dropout-based second view data from the second backbone feature data; and training either one or both of the first neural network model and the second neural network model based on a loss determined based on a combination of any two or more of the first embedded data, the first view data, the second embedded data, the second view data, and an embedded data clustering result.

Abstract: A training method of a neural network, and a recognition method and apparatus using the neural network are disclosed. The recognition method using the neural network includes obtaining a feature vector generated from a hidden layer of the neural network, in response to data being entered to an input layer of the neural network, and determining a reliability of a recognition result for the data using the feature vector and clusters.

Abstract: A training method of a neural network, and a recognition method and apparatus using the neural network are disclosed. The recognition method using the neural network includes obtaining a feature vector generated from a hidden layer of the neural network, in response to data being entered to an input layer of the neural network, and determining a reliability of a recognition result for the data using the feature vector and clusters.

Abstract: A virtual lane generating method and device is provided. The virtual lane generating method includes determining validity of lane detection information extracted from an image in front a vehicle, and generating a virtual lane based on an object included in the image, in response to a determination that the lane detection information is not valid.

Abstract: Disclosed is a method and apparatus for determining an abnormal object, the method including selecting a candidate object from target objects extracted from a two-dimensional (2D) image of a front view captured from a host vehicle, generating a three-dimensional (3D) model of the candidate object, determining, based on the 3D model, whether the candidate object corresponds to an abnormal object that interferes with driving of the host vehicle, and outputting the abnormal object, in response to the candidate object corresponding to the abnormal object.

Abstract: Disclosed is a stereo camera-based autonomous driving method and apparatus, the method including estimating a driving situation of a vehicle, determining a parameter to control a stereo camera width of a stereo camera based on the estimated driving situation, controlling a capturer configured to control arrangement between two cameras of the stereo camera for a first direction based on the determined parameter, and measuring a depth of an object located in the first direction based on two images respectively captured by the two cameras with the controlled arrangement.

Abstract: Disclosed herein are a model training method, a data recognizing method, and a model training apparatus. A model training method includes selecting a teacher model from a plurality of teacher models; receiving, at a student model, input data; and training the student model based on output data of the selected teacher model, the output data corresponding to the input data.

Abstract: A virtual lane generating method and device is provided. The virtual lane generating method includes determining validity of lane detection information extracted from an image in front a vehicle, and generating a virtual lane based on an object included in the image, in response to a determination that the lane detection information is not valid.

Abstract: A training method of a neural network, and a recognition method and apparatus using the neural network are disclosed. The recognition method using the neural network includes obtaining a feature vector generated from a hidden layer of the neural network, in response to data being entered to an input layer of the neural network, and determining a reliability of a recognition result for the data using the feature vector and clusters.

Abstract: Disclosed is a method and apparatus for determining an abnormal object, the method including selecting a candidate object from target objects extracted from a two-dimensional (2D) image of a front view captured from a host vehicle, generating a three-dimensional (3D) model of the candidate object, determining, based on the 3D model, whether the candidate object corresponds to an abnormal object that interferes with driving of the host vehicle, and outputting the abnormal object, in response to the candidate object corresponding to the abnormal object.

Abstract: Disclosed is a stereo camera-based autonomous driving method and apparatus, the method including estimating a driving situation of a vehicle, determining a parameter to control a stereo camera width of a stereo camera based on the estimated driving situation, controlling a capturer configured to control arrangement between two cameras of the stereo camera for a first direction based on the determined parameter, and measuring a depth of an object located in the first direction based on two images respectively captured by the two cameras with the controlled arrangement.

Abstract: Disclosed herein are a model training method, a data recognizing method, and a model training apparatus. A model training method includes selecting a teacher model from a plurality of teacher models; receiving, at a student model, input data; and training the student model based on output data of the selected teacher model, the output data corresponding to the input data.

Abstract: An apparatus and a method to generate an atrial fibrillation prediction model, and an apparatus and a method to predict atrial fibrillation are provided. An atrial fibrillation (AF) prediction model generating apparatus includes a feature extractor configured to extract T-wave features in a predetermined time period from electrocardiogram data and generate a T-wave feature profile based on the extracted features, and a prediction model generator configured to classify the generated T-wave feature profile and generate an AF prediction model using the classified feature profile.

Abstract: An apparatus and a method to generate an atrial fibrillation prediction model, and an apparatus and a method to predict atrial fibrillation are provided. An atrial fibrillation (AF) prediction model generating apparatus includes a feature extractor configured to extract T-wave features in a predetermined time period from electrocardiogram data and generate a T-wave feature profile based on the extracted features, and a prediction model generator configured to classify the generated T-wave feature profile and generate an AF prediction model using the classified feature profile.