Samsung Electronics Co., Ltd.

Minjung Son

Hyun Sung Chang

Donghoon Sagong

Hyong Euk Lee

Inwoo Ha

Keechang Lee

Kyungboo Jung

Heesae Lee

Young Hun Sung

Jiyeon Kim

Wonhee Lee

Minsu Ahn

Seungin Park

Nahyup Kang

Seungeon Kim

Seokhwan Jang

Hwiryong Jung

Hyongeuk Lee

Hyewon Moon

A processor-implemented method with pose estimation includes: determining depth data by sensing a depth of a target scene comprising planes orthogonal to each other; determining normal vectors corresponding to depth points of the depth data; determining orientation data of a device by comparing the normal vectors to orientation candidates; and determining position data of the device based on distances from the device to the planes.

Method and apparatus with pose estimation

A method and apparatus with depth map generation. The method may include generating points for a point cloud by unprojecting multi-view depth maps, of plural views, into a corresponding three-dimensional (3D) space using respective camera parameters corresponding to each view of the multi-view depth maps, extracting feature embedding vectors corresponding to the generated points, generating a two-dimensional (2D) feature map of a set view based on the extracted feature embedding vectors, generating a residual depth map using a refinement network with respect to the 2D feature map, generating a new depth map based on the residual depth map and an initial depth map, of the set view, among the multi-view depth maps.

Method and apparatus with depth map generation

A processor-implemented method includes: determining albedo data in a canonical space and depth data in the canonical space based on input image data including an object, using one or more neural network-based extraction models; generating deformed albedo data and deformed depth data by applying a target shape deformation value respectively to the albedo data and the depth data; generating resultant shaded data by performing shading based on the deformed depth data and a target illumination value; generating intermediate image data based on the resultant shaded data and the deformed albedo data; and generating reconstructed image data from the intermediate image data and the deformed depth data based on a target pose value.

Method and apparatus with image processing and reconstructed image generation

A method with global localization includes: extracting a feature by applying an input image to a first network; estimating a coordinate map corresponding to the input image by applying the extracted feature to a second network; and estimating a pose corresponding to the input image based on the estimated coordinate map, wherein either one or both of the first network and the second network is trained based on either one or both of: a first generative adversarial network (GAN) loss determined based on a first feature extracted by the first network based on a synthetic image determined by three-dimensional (3D) map data and a second feature extracted by the first network based on a real image; and a second GAN loss determined based on a first coordinate map estimated by the second network based on the first feature and a second coordinate map estimated by the second network based on the second feature.

Method and apparatus with global localization

A processor-implemented neural network operating method, the operating method comprising obtaining a neural network pre-trained in a source domain and a first style feature of the source domain, extracting a second style feature of a target domain from received input data of the target domain, using the neural network, performing domain adaptation of the input data, by performing style matching of the input data based on the first style feature of the source domain and the second style feature of the target domain, and processing the style-matched input data, using the neural network.

Apparatus and method with neural network implementation of domain adaptation

An image processing method is provided. The method includes generating appearance information corresponding to a canonical space defined by a cube including a plurality of regular grids in a first input image by implementing a first neural network, generating geometric information corresponding to a standard space in a second input image by implementing a second neural network, deforming the plurality of regular grids based on the generated geometric information, generating standardized image information by applying the appearance information to the plurality of deformed regular grids, and generating an output image based on the standardized image information.

Image processing method and apparatus

A method and apparatus for generating an image and for training an artificial neural network to generate an image are provided. The method of generating an image, including receiving input data comprising conditional information and image information, generating a synthesized image by applying the input data to an image generation neural network configured to maintain geometric information of the image information and to transform the remaining image information based on the conditional information, and outputting the synthesized image.

Computing method and apparatus with image generation

A method and apparatus with feature embedding is provided. The method includes estimating a depth map for each of plural two-dimensional (2D) input images, transforming the depth maps into three-dimensional (3D) information in a point cloud form based on an aggregation of the depth maps, and generating an embedded feature by applying the 3D information to a machine learning model, where the embedded feature includes information about a 3D shape corresponding to a 2D object in the plural 2D input images.

Method and apparatus with feature embedding

Disclosed is a localization method and apparatus that may acquire localization information of a device, generate a first image that includes a directional characteristic corresponding to an object included in an input image, generate a second image in which the object is projected based on the localization information, to map data corresponding to a location of the object, and adjust the localization information based on visual alignment between the first image and the second image.

Localization method and apparatus of displaying virtual object in augmented reality

A method of estimating a position includes: determining a first rotation reference point based on localization information estimated for a target and map data; estimating a second rotation reference point from image data associated with a front view of the target; and correcting the localization information using a rotation parameter calculated based on the first rotation reference point and the second rotation reference point.

Method and apparatus for estimating position

A processor-implemented method using a neural network (NN) includes: receiving input data; and determining information inferred from the input data based on state information about a state in which the NN is activated in response to the input data, wherein an embedding vector generated by encoding the input data using at least a portion of the NN comprises information used to reconstruct a first partial region of the input data with a first accuracy and to reconstruct a second partial region of the input data with a second accuracy, and wherein the first partial region is adaptively determined based on either one or both of the inferred information and the embedding vector.

Method and apparatus with neural network training and inference

A recognizer including a shared encoder to extract a feature of an input image in a source domain and a target domain; and a shared decoder to classify a class of an object included in the input image based on the feature of the input image, in the source domain and the target domain. A set of object classes of the source domain and a set of object classes of the target domain differ from each other.

Recognizer, object recognition method, learning apparatus, and learning method for domain adaptation

Disclosed is a recognition and training method and apparatus. The apparatus may include a processor configured to input data to a neural network, determine corresponding to a multiclass output a mapping function of a first class and a mapping function of a second class, acquire a result of a loss function including a first probability component that changes correspondingly to a function value of the mapping function of the first class and a second probability component that changes contrastingly to a function value of the mapping function of the second class, determine a gradient of loss corresponding to the input data based on the result of the loss function, update a parameter of the neural network based on the determined gradient of loss for generating a trained neural network based on the updated parameter. The apparatus may input other data to the trained neural network, and indicate a recognition result.

Neural network recogntion and training method and apparatus

An apparatus and method to train an image recognition model to accurately estimate a location of a reference point for each class of landmark is disclosed. The apparatus and method use the image recognition model, which is trained based on calculating a class loss and a class-dependent localization loss from training data based on an image recognition model and training the image recognition model using a total loss comprising the class loss and the localization loss.

Growing community of inventors

Suwon-si, South Korea

Minjung Son