Korea Advanced Institute of Science and Technology

Kwanggyoon Seo

Junyong Noh

Sanghun Park

Jung Eun Yoo

Jaedong Kim

Bumki Kim

Hyunggoog Seo

Sihun Cha

Ha Young Chang

Amirsaman Ashtari

Dawon Lee

There is provided an apparatus for generating a 3D object texture map. The apparatus may comprise: a memory; and a processor, wherein the processor is configured to: generate a partial texture image by mapping object information in an input image into a texture space; obtain a sampling image by inputting the partial texture image to a sampler network trained according to a curriculum for selecting at least one of a training partial texture image, an aligned partial image in which the training partial texture image is aligned in the texture space, or an augmented partial image augmented from the aligned partial image; obtain a blending mask and a refined image by inputting the sampling image to a refiner network; and generate a 3D object texture map by blending the sampling image and the refined image based on the blending mask.

Apparatus and method for generating 3D object texture map, and recording medium storing instructions to perform method for generating 3D object texture map

Disclosed is an image processing method and apparatus. The image processing method includes extracting a content latent code and a style latent code for each of a plurality of input images, obtaining a content feature vector by calculating a weighted sum of content latent codes extracted from the input images based on a morphing control parameter, obtaining a style feature vector by calculating a weighted sum of style latent codes extracted from the input images based on the morphing control parameter, and generating a morphing image based on the content feature vector and the style feature vector.

Image processing method and apparatus

A method of training a deep neural network for generating a cinemagraph is disclosed. The method may include preparing a foreground layer input by using an input video, preparing a background layer input by using the input video, providing a foreground layer output and a background layer output from the DNN by inputting, to the DNN, the foreground layer input and the background layer input, providing an output video by synthesizing the foreground layer output with the background layer output, and updating intrinsic parameters of the DNN a plurality of times, based on the input video, the foreground layer input, the foreground layer output, the background layer output, and the output video.

Method of generating cinemagraph based on deep learning

Provided is a method of framing a three dimensional (3D) target object for generation of a virtual camera layout. The method may include analyzing a reference video image to extract a framing rule for at least one reference object in the reference video image, generating a framing rule for at least one 3D target object using the framing rule for the at least one reference object in the reference video image, and using the framing rule for the at least one 3D target object for generation of a virtual camera layout.

Method and apparatus for framing three dimensional target object for generation of virtual camera layout

A method of assisting object registration in a smart home system including a plurality of objects is disclosed. The disclosed method may include allowing a user to select an object from among a plurality of objects, capturing a plurality of pointing gestures by allowing the user to perform the plurality of pointing gestures at the selected object, configuring a pointing ray using each of the plurality of captured pointing gestures, estimating a position of a nearest point extending from the plurality of pointing rays, and registering the estimated position of the nearest point as a position of the selected object.

Growing community of inventors

Daejeon, South Korea

Kwanggyoon Seo