Abstract: Systems, methods, and instrumentalities are disclosed herein for reducing the amortization gap in end-to-end image compression and/or video compression. In examples, a video decoder may obtain an entropy model indication in video data. Based on the entropy model indication, the decoder may determine an entropy model to use for decoding a current picture. The current picture may be decoded based on the determined entropy model. In examples, the entropy model indication may indicate whether to use an updated entropy model or a prior entropy model for decoding the current picture. In examples, the entropy model indication may indicate an updated entropy model or a learned entropy model to use for decoding the current picture.

Abstract: An approach to design a neural network estimates ground contact forces in various areas under each foot from motion capture data at an input. At training time, the input motion capture data is supplemented with synchronized pressure insole data and this extra data is leveraged to optimize neural network weights. At run-time, the optimized network is fed with motion capture data to regress the ground contact forces and predict more accurate ground contact from these forces. The data-driven foot contacts prediction method is learned on a large motion capture database which outperforms traditional heuristic approaches.

Abstract: Methods and apparatuses for encoding or decoding an image or a video based on neural network are provided. In an embodiment, an image is encoded by obtaining a first latent representation of the image in a first latent space, for instance from a Generative Adversarial Network. A second latent representation of the image in a second latent is obtained from the first latent representation and encoded. In an embodiment, the second latent space is obtained from an unfolding of the first latent space based on at least one constraint. In an embodiment, the second latent space is obtained using a neural network. The methods or apparatuses can be used for image editing and/or image or video coding.

Abstract: A method and an apparatus for editing at least two sequences of pictures are provided. Editing the sequences comprises modifying at least one picture of one of the at least two sequences, based on a user input, and updating at least one picture of at least one other sequence of the at least two sequences, with the modification.

Abstract: The present invention generally relates to an apparatus and a method for obtaining a registration error map representing a level of sharpness of an image. Many methods are known which allow determining the position of a camera with respect to an object, based on the knowledge of a 3D model of the object and the intrinsic parameters of the camera. However, regardless of the visual servoing technique used, there is no control in the image space and the object may get out of the camera field of view during servoing. It is proposed to obtain a registration error map relating to an image of the object of interest generated by computing an intersection of a re-focusing surface obtained from a 3D model of said object of interest and a focal stack based on acquired four-dimensional light-field data relating to said object of interest.

Abstract: The present invention generally relates to an apparatus and a method for obtaining a registration error map representing a level of sharpness of an image. Many methods are known which allow determining the position of a camera with respect to an object, based on the knowledge of a 3D model of the object and the intrinsic parameters of the camera. However, regardless of the visual servoing technique used, there is no control in the image space and the object may get out of the camera field of view during servoing. It is proposed to obtain a registration error map relating to an image of the object of interest generated by computing an intersection of a re-focusing surface obtained from a 3D model of said object of interest and a focal stack based on acquired four-dimensional light-field data relating to said object of interest.

Abstract: A method and system is provided for refocusing images captured by a plenoptic camera. In one embodiment the plenoptic camera is in processing with an audio capture device. The method comprises the steps of determining direction of a dominant audio source associated with an image; creating an audio zoom by filtering out all other audio signals except those associated with said dominant audio source; and performing automatic refocusing of said image based on said created audio zoom.

Abstract: The present invention generally relates to an apparatus and a method for obtaining a registration error map representing a level of sharpness of an image. Many methods are known which allow determining the position of a camera with respect to an object, based on the knowledge of a 3D model of the object and the intrinsic parameters of the camera. However, regardless of the visual servoing technique used, there is no control in the image space and the object may get out of the camera field of view during servoing. It is proposed to obtain a registration error map relating to an image of the object of interest generated by computing an intersection of a re-focusing surface obtained from a 3D model of said object of interest and a focal stack based on acquired four-dimensional light-field data relating to said object of interest.

Abstract: A plenoptic camera and associated method is provided. The camera has an array of sensors for generating digital images. The images have associated audio signals. The array of sensors are configured to capture digital images associated with a default spatial coordinate and are also configured to receive control input from a processor to change focus from said default spatial coordinate to a new spatial coordinate based on occurrence of an event at said new spatial coordinate.

Abstract: A method for refocusing, on at least one common point of interest, the rendering of one set of plenoptic video data provided by one plenoptic device belonging to a set of plenoptic devices capturing simultaneously a same scene. According to the present disclosure, said method comprises: obtaining (21) a common 3D reference system used for spatially locating said plenoptic device that has provided said set of plenoptic video data and at least one other device of said set of plenoptic devices, from said at least one common point of interest, determining (22) common refocusing plane parameters in said common 3D reference system, refocusing (23) the rendering of said set of plenoptic video data by converting (231) said common refocusing plane parameters into a rendering refocusing plane of a 3D reference system associated with said plenoptic device.

Abstract: In a context of varying ambient luminosity, a salient idea is to adjust an image by adapting the contrast of an image according to a level of ambient light. Enhancing the contrast is known to make a displayed object more distinguishable, with some impact on the overall image quality. Increasing a level of contrast enhancement as the ambient light increases, advantageously allows to preserve distinguishable displayed objects in high levels of ambient light, despite an overall loss of quality, less perceivable in high levels of ambient light.

Abstract: A method for refocusing, on at least one common point of interest, the rendering of one set of plenoptic video data provided by one plenoptic device belonging to a set of plenoptic devices capturing simultaneously a same scene. According to the present disclosure, said method comprises: obtaining (21) a common 3D reference system used for spatially locating said plenoptic device that has provided said set of plenoptic video data and at least one other device of said set of plenoptic devices, from said at least one common point of interest, determining (22) common refocusing plane parameters in said common 3D reference system, refocusing (23) the rendering of said set of plenoptic video data by converting (231) said common refocusing plane parameters into a rendering refocusing plane of a 3D reference system associated with said plenoptic device.

Abstract: The disclosure relates to a method for transferring a style of a reference visual object to an input visual object. According to an embodiment, the method includes finding a correspondence map assigning to a point in the input visual objet a corresponding point in the reference visual object, the finding of a correspondence map comprising spatially adaptive partitioning of the input visual object into a plurality of regions, the partitioning depending on the reference and input visual objects. The disclosure also relates to corresponding electronic device, computer readable program product and computer readable storage medium.

Abstract: A particular implementation determines color palettes of images by extracting and decomposing color palettes based on the image color content. The decomposition can produce a dictionary matrix, an activation matrix, or both. The dictionary matrix can be used in recoloring an image, either directly or after storing. Another implementation selects a color palette to recolor an image by accessing metadata associated with the image and estimating a scene type based on the metadata and/or other information. Color palettes are retrieved from memory corresponding to the scene type of the image and are used for recoloring the image. Instructions for the implementation can be stored on a non-transitory computer readable medium such that the embodiments can be implemented by one or more processors.

Abstract: A method and system is provided for refocusing images captured by a plenoptic camera. In one embodiment the plenoptic camera is in processing with an audio capture device. The method comprises the steps of determining direction of a dominant audio source associated with an image; creating an audio zoom by filtering out all other audio signals except those associated with said dominant audio source; and performing automatic refocusing of said image based on said created audio zoom.

Abstract: A method, performed by a touch screen device, for delineating an object in an image is disclosed. The method comprises obtaining, from a user input, a band around the object wherein at least one characteristic of the band depends on the presence of artifacts in the neighborhood of the object to delineate. A device for implementing the method and a computer program product are further disclosed.

Abstract: A method of reducing a large amount of media into a sub-group of high quality images in order to capture the diversity of an event. The present invention teaches a method of reducing a plurality of media into clusters in response to time and place. The clusters are further reduced in response to content in said media, including color and facial recognition to transmit still generate highlights. Near duplicate images are then removed from each highlight and then a high quality image is selected from each highlight. The high quality image is selected from each highlight and combined into an event overview to represent the diversity of an event.

Abstract: The present invention generally relates to an apparatus and a method for obtaining a registration error map representing a level of sharpness of an image. Many methods are known which allow determining the position of a camera with respect to an object, based on the knowledge of a 3D model of the object and the intrinsic parameters of the camera. However, regardless of the visual servoing technique used, there is no control in the image space and the object may get out of the camera field of view during servoing. It is proposed to obtain a registration error map relating to an image of the object of interest generated by computing an intersection of a re-focusing surface obtained from a 3D model of said object of interest and a focal stack based on acquired four-dimensional light-field data relating to said object of interest.

Abstract: A method and a device for deblurring a frame of a video are provided. The method includes obtaining (10) a set of neighboring frames of a current frame wherein a global score of sharpness is greater than a sharpness threshold, generating (20) of a local blur map, delivering the local blur map, performing (30) a local warping of at least one frame of the set and of the local blur map as a function of a local motion estimation between the current frame and the at least one frame of the set, delivering at least one locally warped frame and an associated locally warped blur map and performing (40) a weighted aggregation of a part of the at least one locally warped frame and a corresponding part the current frame, based on the at least one locally warped blur map and the local blur map of the current frame.

Abstract: One general aspect for motion-based video tonal stabilization uses a keyframe and motion estimation techniques to determine the level of spatial correspondence between input images and the keyframe. When the level of spatial correspondence is high, tonal stabilization is performed through regression and power law tonal transformation to minimize the color differences between images caused by automatic camera parameters without a priori knowledge of the camera model. Tonal error accumulation is reduced by using long-term tonal propagation.