Samsung Electronics Co., Ltd.

Anton Sergeevich Konushin

Danila Dmitrievich Rukhovich

Do Kyoon Kim

Kee Chang Lee

Hwa Sup Lim

Olga Barinova

Radoslaw Pawel Chmielewski

Sun Hyuck Chae

Piotr Kacper Wierzba

Mikhail Viktorovich Romanov

Grzegorz Kisala

Anton Anatolyevich Yakubenko

Alexander Georgievich Limonov

Danil Fanilievich Galeev

Ilia Igorevich Belikov

Vadim Sergeevich Konushin

Anna Borisovna Vorontsova

Konstantin Sergeevich Sofiyuk

Kamil Przybyla

A method of 3D object detection using an object detection neural network includes: receiving one or more monocular images; extracting 2D feature maps from each one of the one or more monocular images by passing the one or more monocular images through a 2D feature extracting part of the object detection neural network, generating an averaged 3D voxel volume based on the 2D feature maps, extracting a 2D representation of 3D feature maps from the averaged 3D voxel volume by passing the averaged 3D voxel volume through an encoder of a 3D feature extracting part of the object detection neural network, and performing 3D object detection as 2D object detection in a Bird's Eye View (BEV) plane, the 2D object detection in the BEV plane being performed by passing the 2D representation of 3D feature maps through thane outdoor object detecting part of the object detection neural network.

Method and electronic device for 3D object detection using neural networks

The present invention relates to methods for visual display of images of real shelvings with products for analysis of shelving contents. There are provided methods for constructing a shelving front panorama. A method comprises: capturing color image frames of the shelving, displaying the shelving and its contents; reconstructing a shelving 3D model based of depth data and capturing device position data for each frame; determining a projection plane corresponding to the shelving front edge; stitching the color image frames of the shelving by projective transformation of each color image frame of the shelving onto the projection plane. The resulting shelving front panorama displays the shelving and its contents. The disclosure eliminates the need for observing frontal orientation of each shelving image frame when capturing images of the shelving; the capturing can be made along an arbitrary path.

Method of constructing front panorama of shelving from arbitrary series of frames based on shelving 3D model

A method and apparatus for joint image and per-pixel annotation synthesis with a generative adversarial network (GAN) are provided. The method includes: by inputting data to a generative adversarial network (GAN), obtaining a first image from the GAN; inputting, to a decoder, a first feature value that is obtained from at least one intermediate layer of the GAN according to the inputting of the data to the GAN; and obtaining a first semantic segmentation mask from the decoder according to the inputting of the first feature value to the decoder.

Teaching GAN (generative adversarial networks) to generate per-pixel annotation

A method of determining a ground line from an input image. The method includes determining a plurality of ground line candidates from the image, determining a certain band a central-line of which is a boundary between a G region and other regions in a Ground Building Sky (GBS) map of the image, and determining the ground line of the image by selecting a ground line candidate, among the plurality of ground line candidates, having the greatest extent crossing the certain band.

Method for determining ground line

Growing community of inventors

Moscow, Russia

Anton Sergeevich Konushin