Abstract: The invention relates to a method and a device for converting an original image sequence (SIO) whose luminance values belong to a given dynamic range, said method comprising a step of image-by-image dynamic range conversion during which a dynamic range conversion operator (TMO) is applied to the luminance component (Lw(i)) of each image of the original image sequence (SIO) to obtain a modified luminance component whose values belong to a lower dynamic range than that of the original image sequence (SIO). The method is characterized in that it also comprises a step of correction of the dynamic range during which the dynamic range of the values of the luminance component thus modified is reduced by a value calculated from a global characteristic of the original image sequence.

Abstract: The present invention generally relates to a method for predicting a block of pixels from at least one patch comprising a block of pixels and a causal neighborhood around this block of pixels. The method comprises the following steps: determining a mapping of a causal neighborhood, around that block of pixels to be predicted, on the block of pixels to be predicted in order that the block of pixels of each patch is best predicted by mapping the neighborhood of that patch on the block of pixels of that patch, and predicting the block of pixels from a prediction block computed by applying the determined mapping on the neighborhood of the block of pixels to predict.

Abstract: A method for enhancing quality of an image comprises deriving from an initial low-resolution image (LRI) an initial high-resolution image (IHRI) by upsampling (S2, S3), providing (S4), based on the initial low-resolution image (LRI), at least one downsampled filtered image (AIij, AIi?j?) with lower resolution, providing (S6), based on the initial low-resolution image (LRI), an unfiltered image (LRI, UIi?j?) having same resolution as the downsampled filtered image, selecting a patch (PI) from the initial high-resolution image (IHRI), finding (S8) filial patches (PA k) similar to the selected patch in the downsampled filtered image (AIij, AIi?j?), finding, in the unfiltered image (LRI, UIi?j?), parent patches (PPk) locally associated to the filial patches (PAk), and linearly combining (S14) the parent patches (PPk) to form an enhanced quality patch (EP).

Abstract: A method of encoding or decoding at least part of a high dynamic range image, the image being defined in a perceptual space of high dynamic range having a luminance component and a color difference metric, the method comprising for a block of the at least part of the image: converting reference samples for prediction of the block into the LDR space of the block to be predicted, the reference samples having been previously reconstructed in their associated LDR space, and predicting the block using the converted reference samples, for encoding or decoding of the block using an encoding or decoding technique applicable to an LDR image.

Abstract: A method for encoding a matrix of views obtained from light-field data is disclosed. The method includes obtaining information representative of a type of encoding and encoding such information, wherein the type of encoding corresponding to an epipolar plane image encoding using at least one part of epipolar plane image, or said type of encoding corresponding to a view by view encoding according to a views' encoding order, and encoding the group of views in said data stream responsive to said information representative of a type of encoding, wherein, in the case where said type of encoding is an epipolar plane image encoding. The group of views is encoded by obtaining at least one part of an epipolar plane image, and encoding said at least one part of an epipolar plane image.

Abstract: A method for constructing an epitome from an image divided into non overlapping blocks is disclosed. The method comprises: determining, for each block, similar patches in the image, a similar patch being a patch with similar content; constructing at least one epitome chart for the picture from the similar patches; wherein determining, for each block, similar patches in the image comprises: a) determining, for each current block, similar blocks in the image, a similar block being a block with content similar to the content of the current block; b) determining, for one current block and for the similar blocks determined for the current block, similar patches in the image, a similar patch being a patch with content similar to the content of the current block; c) repeating step b) for a next current block for which no similar patch is determined until at least one similar patch is determined for each block in the image.

Abstract: Super-resolution refers to a process of recovering the missing high-frequency details of a given low-resolution image. Known single image SR algorithms are often computationally intractable or unusable for most of the practical applications. The invention relates to a method for performing hierarchical super-resolution based on self content neighboring patches information is based on pyramidal decomposition. The intrinsic geometric property of an input LR patch neighborhood is obtained from the input LR patch and its K nearest neighbors taken from different down-scaled versions of the LR image.

Abstract: A method and device for processing LDR images of a video sequence to improve image quality. The method comprises temporally decomposing successive HDR frames of a video sequence and the corresponding LDR frames and performing a comparison between the HDR and LDR frequency sub-bands. A current LDR image can then be modified on the basis of a comparison between the frequency sub-bands.

Abstract: The following steps are carried out in the inventive method: estimation of movement between a current moment corresponding to the current image and a reference moment corresponding to a reference image in order to provide a movement vector, an entropic coding for a high spatial frequency signal relating in the current image, taking into account a temporal context based on the estimation of a movement, one of the parameters used to calculate the temporal context being the parity of a component of the movement vector. Applications are related to video compression with temporal prediction.

Abstract: A method and an apparatus for de-noising an image, and in particular, de-noising an image using video epitome based on a source video image. An embodiment of the present principles provides a method of processing an image in a video, comprising: decoding an encoded version of the image to generate a decoded version of the image; and generating a de-noised version of the image using the decoded version of the image and a video image epitome which is a texture epitome associated with the image, wherein the video image epitome was extracted from a source version of the image, wherein the generating comprises: de-noising a current patch using corresponding patches located in the video image epitome that correspond to at least one of a plurality of nearest neighbor patches.

Abstract: The invention is made in the field of coding of images of high dynamic range. The invention is based on the concept of Frame Compatible format. The idea is to transport, in a frame, down-sampled LDR content together with additional information allowing reconstructing HDR content from the LDR content. Thus, it is proposed a method of encoding an HDR image of high dynamic range according to claim 1. Said method comprises down-sampling (DWN) an LDR image and additional data, the LDR image providing a lower dynamic range depiction of the HDR image content and the additional data allowing for reconstructing the HDR image from the LDR image.

Abstract: The invention relates to a method for predicting a pixel block of an image using a weighted sum of pixel blocks belonging to patches of a dictionary from a set of candidate patches, each patch being formed of a pixel block of an image and a causal neighborhood of this pixel block. The method is wherein a subset of candidate patches is obtained from the set of candidate patches and the dictionary is formed of a patch of the subset of candidate patches, called the first patch, and at least one other patch of said set of candidate patches, called the second patch.

Abstract: The invention relates to a method and device for quantising the floating value of a pixel of an image by rounding either to a lesser whole number, to a greater whole number, or to the whole number closest to this floating value. The method is characterised in that the selection of rounding this floating value is determined based on a test value.

Abstract: To encode High Dynamic Range (HDR) images, the HDR images can be converted to Low Dynamic Range (LDR) images through tone mapping operation, and the LDR images can be encoded with an LDR encoder. The present principles formulates a rate distortion minimization problem when designing the tone mapping curve. In particular, the tone mapping curve is formulated as a function of the probability distribution function of the HDR images to be encoded and a Lagrangian multiplier that depends on encoding parameters. At the decoder, based on the parameters indicative of the tone mapping function, an inverse tone mapping function can be derived to reconstruct HDR images from decoded LDR images.

Abstract: The invention relates to a method for processing a video sequence formed by at least two video images, said method comprising: obtaining said at least two video images using a first capture module with a first field of view, each video image representing a same scene captured at different instants or from different points of view, wherein said method comprises: determining luminance information from at least one overall image obtained by a second capture module with a second field of view greater than said first field of view, said overall image representing the scene captured at different instants or from different points of view, down-converting a dynamic range of said video images taking into account said luminance information.

Abstract: The invention relates to image block encoding and decoding. More precisely, the invention relates to block prediction using reference blocks of the image. It is proposed a method which comprises using processing means for using at least the block for determining one reference block in an already encoded part of the image, encoding and reconstructing a part of the image, using at least the block for determining one reference block in the reconstructed part and a vector referencing the one reference block, using the one reference block for determining at least one further reference block in the reconstructed part and encoding the determined vector and a residual of the block with respect to a combination of the one and the at least one further reference block. Using the block for determination of the one reference block allows for selecting it such that it resembles the block and correlation with the block is maintained.

Abstract: A method for generating at least one image with a first dynamic range, from an image with a second dynamic range, which is lower than the first dynamic range is described. The method includes obtaining an epitome of the image with a first dynamic range, called a first epitome. Thereafter, the image with a first dynamic range is generated from the image with a second dynamic range and the first epitome.

Abstract: The present invention concerns a method and apparatus for building an estimate (?) of an original image (Y) from a low-quality version (Yl) of the original image and an epitome (Eh) calculated from an image.

Abstract: The present principles provide a method and an encoder for encoding an image frame. The method includes generating, with a processor, for each block in the image frame a prediction block based on at least a prediction method; generating a scaling factor and an offset for adjusting the prediction block; and transforming the prediction block into an adjusted prediction block based on the scaling factor and the offset. The encoder includes a processor configured to, for each block in the image frame, generate a prediction block based on at least a prediction method and generate a scaling factor and an offset for adjusting the prediction block and to transform the prediction block into an adjusted prediction block based on the scaling factor and the offset. The scaling factor and offset minimize error between the block and the adjusted prediction block.

Abstract: A method for coding a current block using a parametric motion model whose parameters are defined from the motion data of the causal neighborhood is disclosed. Specifically, the method for coding the current block comprises the following steps: determine a prediction block of the current block from a motion model defined by a set of parameters, the model associating a motion vector pointing at a pixel of a prediction block with each pixel of the current block. calculate a residual block between the current block and the prediction block, and code the residual block. According to the invention, the parameters of the motion model are determined from the motion data associated with the causal neighborhood blocks of the current block. Therefore, they can be deducted from parts of images already coded and do not need to be incorporated in the image data stream.