Abstract: A method and system are provided for processing image content. The method comprises receiving information about a content image captured at least by one camera. The content includes multi-view representation of an image including both distorted and undistorted areas. The camera parameters and image parameters are then obtained and used to determine to which areas are undistorted and which areas are distorted in said image. This is used to calculate depth map of the image using the determined undistorted and distorted information. A final stereoscopic image is then rendered that uses the distorted and undistorted areas and calculation of depth map.

Abstract: Methods and devices for generated multiplane images from a representation of a 3D scene are disclosed. When an MPI is generated by the system, an image of the accumulation of the alpha values is computed. A convolutional networked is trained with a penalization function that penalizes MPIs with high values of accumulated alpha according to a parameter selected or trained to reduce the redundancy of information between layers of the MPI. When generating a MPI from a subset of the views of a new captured 3D scene, a first MPI is generated with the trained CNN with the associated parameter. An error is calculated between the other views and corresponding synthesized with the first MPI. This error is used to modify the parameter. Then, a second MPI is generated with the trained CNN with the new parameter.

Abstract: In one embodiment of the disclosure, it is proposed a method for encoding a matrix of image views obtained from data acquired by a plenoptic camera. The method is remarkable in that it comprises: obtaining at least one reference image view from a combination of at least two image views from said matrix of image views; encoding image views that are different from said at least one reference image view, based on said at least one reference image view.

Abstract: A method for reducing the parameters defining an acquired light field ray which enables only the colour associated with the light field ray to be stored instead of 4 light field co-ordinates (x,y,i,j) and its associated colour.

Abstract: The invention relates to layered depth data. In multi-view images, there is a large amount of redundancy between images. Layered Depth Video format is a well-known formatting solution for formatting multi-view images which reduces the amount of redundant information between images. In LDV, a reference central image is selected and information brought by other images of the multi-view images that are mainly regions occluded in the central image are provided. However, LDV format contains a single horizontal occlusion layer, and thus fails rendering viewpoints that uncover multiple layers dis-occlusions. The invention uses light filed content which offers disparities in every directions and enables a change in viewpoint in a plurality of directions distinct from the viewing direction of the considered image enabling to render viewpoints that may uncover multiple layer dis-occlusions which may occurs with complex scenes viewed with wide inter-axial distance.

Abstract: The present disclosure relates to method and apparatus for the estimation of an accurate multiplane image from a set of posed images. The accuracy of an MPI representation is measured by the performance of the views generated by it through inverse homography sampling and alpha compositing. A variable number of posed images at a possibly different resolution representative of scenes with arbitrary geometry are used as input of the system. A first level of three modules, possibly trained based on deep learning methods, is built for extracting view and scene features and color scores. Numerous levels of the same three modules may be built and organized in a structure where a level deals with down-scaled version of the posed images.

Abstract: An apparatus and a method are provided for image processing. In one embodiment, the method comprises accessing a plurality of images captured by at least a reference camera, wherein the images represent a plurality of views corresponding to said same scene. A plurality of plane sweep volume (PSV) slices are then generated from said images and computing for each slice a flow map from at least the reference camera calibration parameter and this flow map a previous slice of the plane sweep volume is generated.

Abstract: Predicting a component of a current pixel belonging to a current sub-aperture image in a matrix of sub-aperture images captured by a sensor of a type I plenoptic camera can involve, first, determining a location on the sensor based on: a distance from an exit pupil of a main lens of the camera to a micro-lens array of the camera; a focal length of the main lens; a focal length of the micro-lenses of the micro-lens array; and a set of parameters of a model of the camera allowing for a derivation of a two-plane parameterization describing the field of rays corresponding to the pixels of the sensor; and, second, predicting the component based on one reference pixel belonging to a reference sub-aperture image in the matrix and located on the sensor in a neighborhood of the location.

Abstract: A method and system are provided for generating an image from each camera array in a camera array matrix. In one embodiment, the method comprises increasing redundancy between to be captured images from the camera matrix by rotating direction of any cameras disposed in upper and lower rows of said matrix by a 90 degree angle around the roll axis. Then any cameras disposed at the corners of the matrix are rotated in an angle that is less than 90 degrees around the roll axis. Subsequently, the location of central cameras are determined and analysed so that they can be rotated and disposed in a manner that provides both horizontal and vertical compensation for any redundancies.

Abstract: A device includes at least one first sensor and at least one second sensor for capturing first image data, the at least one first and second sensors being rectangular and arranged orthogonal to the at least one first rectangular sensor, and at least one hardware processor configured to cause the at least one first and second sensors, respectively, the first image data and the second image data at least substantially simultaneously, and at least one of display simultaneously data from the first image data and data from the second image data as a cross-shaped image or store together data from the first image data and data from the second image data as a cross-shaped image. The resulting first and second image data can be stored in a single file in memory. The at least one hardware processor can be processed to remove redundancies between the image data. The device can also extract from the first and second image data, image data corresponding to a rectangle parallel with an horizon.

Abstract: A method and system are provided for processing image content. In one embodiment the method comprises receiving a plurality of captured contents showing same scene as captured by one or more cameras having a different focal length and depth maps and generating a consensus cube by obtaining depth map estimations from said received contents. The visibility of different objects in then analysed to create a soft visibility cube that provides visibility information about each content. A color cube is then generated by using information from the consensus and soft visibility cube. The color cube is then used to combine different received contents and generate a single image for the plurality of contents received.

Abstract: Predicting a component of a current pixel belonging to a current sub-aperture image in a matrix of sub-aperture images captured by a sensor of a type I plenoptic camera can involve, first, determining a location on the sensor based on: a distance from an exit pupil of a main lens of the camera to a micro-lens array of the camera; a focal length of the main lens; a focal length of the micro-lenses of the micro-lens array; and a set of parameters of a model of the camera allowing for a derivation of a two-plane parameterization describing the field of rays corresponding to the pixels of the sensor; and, second, predicting the component based on one reference pixel belonging to a reference sub-aperture image in the matrix and located on the sensor in a neighborhood of the location.

Abstract: A method for decoding or encoding includes obtaining views parameters for a set of views comprising at least one reference view and a current view of a multi-views video content wherein each view comprises a texture layer and a depth layer. For at least one couple of a reference view and the current view of the set of views, an intermediate prediction image applying a forward projection method to pixels of the reference view is generated to project these pixels from a camera coordinates system of the reference view to a camera coordinates system of the current view, the prediction image comprising information allowing reconstructing image data. At least one final prediction image obtained from at least one intermediate prediction image is stored in a buffer of reconstructed images of the current view. A current image of the current view from the images stored in said buffer is reconstructed, said buffer comprising said at least one final prediction image.

Abstract: A method and system are provided for processing image content. The method comprises receiving information about a content image captured at least by one camera. The content includes multi-view representation of an image including both distorted and undistorted areas. The camera parameters and image parameters are then obtained and used to determine to which areas are undistorted and which areas are distorted in said image. This is used to calculate depth map of the image using the determined undistorted and distorted information. A final stereoscopic image is then rendered that uses the distorted and undistorted areas and calculation of depth map.

Abstract: For multi-view video content represented in the MVD (Multi-view+Depth) format, the depth maps may be processed to improve the coherency therebetween. In one implementation, to process a target view based on an input view, pixels of the input view are first projected into the world coordinate system, then into the target view to form a projected view. The texture of the projected view and the texture of the target view are compared. If the difference at a pixel is small, then the depth of the target view at that pixel is adjusted, for example, replaced by the corresponding depth of the projected view. When the multi-view video content is encoded and decoded in a system, depth map processing may be applied in the pre-processing and post-processing modules to improve video compression efficiency and the rendering quality.

Abstract: A method is proposed for estimating a depth for pixels in a matrix of M images. Such method comprises, at least for one set of N images among the M images, 2<N?M, a process comprising: —determining depth maps for the images in the set of N images delivering a set of N depth maps; —for at least one current pixel for which a depth has not yet been estimated: —deciding if a candidate depth corresponding to a depth value in the set of N depth maps is consistent or not with the other depth map(s) of the set of N depth maps; —selecting the candidate depth as being the estimated depth for the current pixel if the candidate depth is decided as consistent. The process is enforced iteratively with a new N value which is lower than the previous N value used in the previous iteration of the process.

Abstract: The invention relates to layered depth data. In multi-view images, there is a large amount of redundancy between images. Layered Depth Video format is a well-known formatting solution for formatting multi-view images which reduces the amount of redundant information between images. In LDV, a reference central image is selected and information brought by other images of the multi-view images that are mainly regions occluded in the central image are provided. However, LDV format contains a single horizontal occlusion layer, and thus fails rendering viewpoints that uncover multiple layers dis-occlusions. The invention uses light filed content which offers disparities in every directions and enables a change in viewpoint in a plurality of directions distinct from the viewing direction of the considered image enabling to render viewpoints that may uncover multiple layer dis-occlusions which may occurs with complex scenes viewed with wide inter-axial distance.

Abstract: The invention relates to layered depth data. In multi-view images, there is a large amount of redundancy between images. Layered Depth Video format is a well-known formatting solution for formatting multi-view images which reduces the amount of redundant information between images. In LDV, a reference central image is selected and information brought by other images of the multi-view images that are mainly regions occluded in the central image are provided. However, LDV format contains a single horizontal occlusion layer, and thus fails rendering viewpoints that uncover multiple layers dis-occlusions. The invention uses light filed content which offers disparities in every directions and enables a change in viewpoint in a plurality of directions distinct from the viewing direction of the considered image enabling to render viewpoints that may uncover multiple layer dis-occlusions which may occurs with complex scenes viewed with wide inter-axial distance.

Abstract: A method is proposed for estimating a depth for pixels in a matrix of M images. Such method comprises, at least for one set of N images among the M images, 2?N?M, a process comprising: —determining depth maps for the images in the set of N images delivering a set of N depth maps; —for at least one current pixel for which a depth has not yet been estimated: —deciding if a candidate depth corresponding to a depth value in the set of N depth maps is consistent or not with the other depth map(s) of the set of N depth maps; —selecting the candidate depth as being the estimated depth for the current pixel if the candidate depth is decided as consistent. The process is enforced iteratively with a new N value which is lower than the previous N value used in the previous iteration of the process.

Abstract: There are several types of plenoptic devices having their proprietary file format. At present there is no standard supporting the acquisition and transmission of multidimensional information for an exhaustive over-view of the different parameters upon which a light-field depends. As such acquired light-field data for different cameras have a diversity of formats. The notion of pixel beam, which represents a volume occupied by a set of rays of light in an object space of an optical system of a camera is thus introduced. The method according to the invention enables to provide data representative of a collection of pixel beams describing a first optical system that is agnostic since these data are obtained by imaging the collection of pixel beams through a second optical system. Such data representative of a collection of pixel beams enable the generation of parametrized output images from which post-processing.