Politechnika Poznanska

Electronics and Telecommunications Research Institute

Poznan University of Technology

Huawei Technologies Co., Limited

Marek Domanski

Krzysztof Wegner

Olgierd Stankiewicz

Krzysztof Klimaszewski

Jakub Stankowski

Tomasz Grajek

Adrian Dziembowski

Dawid Mieloch

Damian Karwowski

Gwang Soon Lee

Jun Young Jeong

Robert Ratajczak

Gwang-Soon Lee

Jacek Konieczny

Jun-Young Jeong

Adam Luczak

Jakub Siast

Maciej Kurc

Kug Jin Yun

Hong Chang Shin

Jin-Hwan Lee

Slawomir Mackowiak

Adam Łuczak

Adam Grzelka

Blazej Szydelko

Disclosed herein are an immersive video encoding method and an immersive video decoding method. The immersive video encoding method includes setting basic view images among input images corresponding to multiple views, generating an atlas image using the basic view images, performing encoding on the atlas image, and generating metadata about an encoded atlas image, wherein the atlas image includes a texture atlas or a depth atlas.

Immersive video encoding method and immersive video decoding method

Disclosed herein are an immersive video decoding method and an immersive video encoding method. The immersive video decoding method includes partitioning a current picture into multiple main blocks, determining whether each main block is to be partitioned into multiple sub-blocks, and when it is determined that each main block is to be partitioned into multiple sub-blocks, determining a partitioning type of the corresponding main block. Here, the corresponding main block may be partitioned into four or two sub-blocks depending on the partitioning type. When the corresponding main block is partitioned into multiple sub-blocks, whether depth information is to be updated for each of the sub-blocks generated by partitioning the main block may be determined.

Immersive video decoding method and immersive video encoding method

A method for processing an immersive video includes: performing pruning for view images; generating an atlas by packing a patch that is extracted as a result of the pruning; deriving an offset for the patch that is comprised in the atlas; and correcting pixel values in the patch by using the derived offset.

Method for processing immersive video and method for producing immersive video

An immersive video processing method according to the present invention includes: classifying view videos into a base video and an additional video; performing pruning for the view videos by referring to a result of the classification; generating an atlas based on a result of the pruning; determining a depth parameter of each view in the atlas; and encoding information indicating whether or not updating of the depth parameter is needed, based on whether or not the depth parameter of each view in the atlas is identical as in a previous atlas.

Method for processing immersive video and method for producing immversive video

Disclosed herein is an image encoding/decoding method and apparatus for virtual view synthesis. The image decoding for virtual view synthesis may include decoding texture information and depth information of at least one or more basic view images and at least one or more additional view images from a bit stream and synthesizing a virtual view on the basis of the texture information and the depth information, wherein the basic view image and the additional view image comprise a non-empty region and an empty region, and wherein the synthesizing of the virtual view comprises determining the non-empty region through a specific value in the depth information and a threshold and synthesizing the virtual view by using the determined non-empty region.

Method and apparatus for encoding/decoding image for virtual view synthesis

The present disclosure can provide an image processing apparatus. In some embodiments, the image processing apparatus in accordance with the disclosure comprises memory configured to store a previously selected depth information value (d), (d; d) having been selected for a previously processed fragment and signal processing logic (). The signal processing logic is configured to calculate a similarity measure for the current fragment based on a depth information value candidate defining a reference fragment candidate of a digital reference image, compare the depth information value candidate with the previously selected depth information value (d), apply a weighting function to the calculated similarity measure to obtain a weighted similarity measure for the current fragment based on the comparison, and determine whether to select the depth information value candidate as the depth information value (d) for the current fragment based on the weighted similarity measure of the depth information value candidate.

Image processing apparatus and method

A method for disoccluded region coding in multiview video data stream by an entropy encoder, the method comprising the steps of: coding a block of a base view; storing state and estimated probabilities in contexts models of the entropy encoder in a context storage module with a reference identifying the block of the base view; repeating the aforementioned steps of coding and storing for every block of the base view of the multiview video data stream; starting coding of a disoccluded region and dividing, into blocks, disoccluded area of a side view associated with the base view; determining, for neighboring blocks, of currently coded block, that have not been in the disoccluded area, a corresponding block in the base view, using a block correspondence database; when such a correspondence is determined, reading a previously stored state and estimated probabilities in context models of the entropy encoder for the corresponding block; copying all coding modes from the corresponding block to the neighboring block; applying state and estimated probabilities in context models to the entropy encoder of the block that is adjacent to the currently processed block; encoding, by the entropy encoder, the selected block of the side view; and repeating steps (e) to (i) for all blocks of the side view associated with the base view processed in steps (a) to (b).

System and a method for disocluded region coding in a multiview video data stream

A method for predictive encoding of a depth map, includes the steps of receiving a depth map; dividing the depth map to blocks; performing quad-tree decomposition of the depth map to sub-blocks; approximating each of the sub-blocks by a plane, wherein the plane is associated with three plane points; and determining an order of coding of the sub-blocks; and providing an encoded depth map in a form of a set of planes defined by parameters. The method includes, while encoding consecutive sub-blocks: calculating predictors for the plane points of the currently-encoded sub-block (C); calculating differences between the actual values of the plane points and the values of the corresponding predictors; and providing the parameters of the plane for the currently-encoded sub-block (C) in form of the differences.

Method for predictive coding of depth maps with plane points

A method for video object tracking using stereoscopic images with depth maps. The method comprises calculating a first axis histogram for each depth map and applying: a first object detection method to track objects based on at least one of: the contents of the images and the depth maps; and a second object detection method to track objects based on the contents of the histograms of the depth maps. The locations of tracked objects are determined based on comparison of the results of the first object detection method and the second object detection method.

System and a method for tracking objects

A computer-implemented system for generating a depth map for a pair of stereoscopic images, the system comprising a lower level processing arrangement and a higher level processing arrangement. The lower level processing arrangement comprises a search range estimation module receiving a reliability map and configured to determine a range of considered disparities in a higher level of processing wherein a decision regarding such range is taken based on the reliability map. The higher level processing arrangement comprises a depth map generation with reliability module receiving output of an image sections matching module configured to output a depth map and a reliability map wherein the reliability map comprises data on reliability, with which a disparity value has been determined for each point of the pair of stereoscopic images.

System and a method for generating a depth map

A method for depth-image-based rendering, the method comprising the steps of: obtaining a first reference view; obtaining a depth map for the first reference view; obtaining a second reference view; obtaining a depth map for the second reference view; the method further comprising the steps of extracting noise present in the first and the second reference views; denoising the first and the second reference views and, based on the denoised first and second reference views, rendering an output view using depth-image-based rendering; adding the extracted noise to the output view.

System and a method for depth-image-based rendering

A method for estimation of dimensions of an object present on images obtained from a set of N cameras and representing the same scene from different viewpoints, based on a 3D object model. The method comprises determining () a 3DMM model. It further comprises matching () the 3DMM model to each of N images of the set in order to determine a set of parameters h, &#x3c1;, &#x3bb;, wherein hdescribes the orientation of the matched model with respect to the i-th image; &#x3c1;describes the 3D shape of the model matched to the i-th image and &#x3bb;describes the appearance/texture of the model matched to the i-th image. The method comprises calculating () relative positions of the cameras, from which the images were collected, based on 3D transformations Hof the matched 3DMM model from the i-th image to the matched 3DMM model from the j-th image; and calculating () the total matching error Efor the model as a sum of matching errors of a model described by parameters h, &#x3c1;, &#x3bb;to individual images:  The method further comprises estimating () the dimensions of the object based on locations of characteristic points from vector s, calculated as: whereinis the vector of the average 3DMM model shape, &#x3c1;are elements of vector &#x3c1;, and &#x15d;are eigenvalue vectors of the 3DMM model.

System and method for object dimension estimation using 3D models

A method for estimating a size of an object present on two images representing the same scene from different points of view, the method comprising: determining an AAM image model; matching the AAM image model to the images to find a set of parameters describing the shape of the model (p) and the appearance of the model (&#x3bb;) for which minimal matching errors for both images occur; and estimating the size of the object based on the differences in locations of points of the AAM image model in the images. The image model is matched to both images mutually, i.e. by using information from both images.

System and method for object dimension estimation

An image coding method. The method includes encoding a first image, the first image having a reduced spatial resolution relative to an input image, wherein values of the first image depend on energy of corresponding areas of the input image, and encoding a first shape information of the shape of a spatial spectrum of the input image. The method further includes reconstructing a second image from the encoded first image, reconstructing a second shape information from the encoded first shape information, and generating a noise image having an intensity described by the second image and having a spectrum shape described by the second shape information.

Growing community of inventors

Poznan, Poland

Marek Domanski