Samsung Electronics Co., Ltd.

Emc IP Holding Company LLC

Leonid Levkovich-Maslyuk

Mahn-jin Han

In-Kyu Park

Alexey Victorovich Ignatenko

David Stephen Reiner

Nihar Nanda

Alexander Olegovich Zhirkov

Anton Konouchine

Joo-yeon Han

Andrey Abramov

An apparatus in one embodiment comprises a processing platform implementing a data set discovery engine. The data set discovery engine comprises a data set indexer configured to generate similarity indexes for a plurality of data sets, and a relativistic retriever coupled to the data set indexer and configured to obtain a suitability template for a query and to execute the query against one or more of the similarity indexes based at least in part on the suitability template. A given one of the similarity indexes comprises at least first and second auxiliary information generated from respective ones of at least first and second different similarity measures of a plurality of different similarity measures. The first and second similarity measures comprise selected ones of the plurality of different similarity measures that are supported by the data set discovery engine with the supported similarity measures comprising both frequency-based and non-frequency-based similarity measures.

Data set discovery engine comprising relativistic retriever

A family of node structures for representing 3-dimensional objects using depth image are provided. These node structures can be adopted into MPEG-4 AFX for conventional polygonal 3D representations. Main formats of the family are DepthImage, PointTexture and OctreeImage. DepthImage represents an object by a union of its reference images and corresponding depth maps. PointTexture represents the object as a set of colored points parameterized by projection onto a regular 2D grid. OctreeImage converts the same data into hierarchical octree-structured voxel model, set of compact reference images and a tree of voxel-image correspondence indices. DepthImage and OctreeImage have animated versions, where reference images are replaced by videostreams. DIBR formats are very convenient for 3D model construction from 3D range-scanning and multiple source video data. MPEG-4 framework allows construction of a wide variety of representations from the main DIBR formats, providing flexible tools for effective work with 3D models. Compression of the DIBR formats is achieved by application of image (video) compression techniques to depth maps and reference images (videostreams).

Node structure for representing 3-dimensional objects using depth image

A method and apparatus removing redundant data of a simple texture model, using an extended depth image, and an image-based editing method and apparatus. The method involves calculating normal vectors of pixels projected into a three-dimensional (3D) space by using depth images among a plurality of simple texture images, the plurality of simple texture images rendering different aspects of a 3D model; calculating reliabilities of the pixels of each of the simple texture images by using the calculated normal vectors; and comparing the reliabilities of the pixels that originate from different simple texture images but render the same portion of the 3D model and removing the pixels having low reliabilities from the different simple texture images from which they respectively originate.

Image-based rendering and editing method and apparatus

Growing community of inventors

Moscow, Russia

Leonid Levkovich-Maslyuk