Abstract: Apparatus for processing image data associated with at least one input image, including a convolutional neural network, CNN,-based encoder configured to provide a plurality of hierarchical feature maps based on the image data, a decoder configured to provide output data based on the plurality of feature maps, wherein the decoder includes a convolutional long short-term memory, Conv-LSTM, module configured to sequentially process at least some of the plurality of hierarchical feature maps.

Abstract: An apparatus configured to, in respect of virtual reality content comprising video imagery configured to provide a virtual reality space wherein a virtual reality view presented to a user provides for viewing of the VR space; based on a comment made by and a virtual location of a commenting-user in the virtual reality space when the comment was made; provide for determination of a point of interest in the virtual reality space, the point of interest identified based on, at least, the virtual location of the commenting-user when the comment was made and semantic analysis of the comment to identify the point of interest surrounding the virtual location to which the comment refers, the point of interest associated with the comment thereby enabling the comment to be overlaid over the virtual reality view of the video imagery.

Abstract: There is provided an apparatus comprising means for performing: training a neural network by applying an optimization loss function, wherein the optimization loss function considers empirical errors and model redundancy (210); pruning a trained neural network by removing one or more filters that have insignificant contributions from a set of filters (220); and providing the pruned neural network for transmission (230).

Abstract: Apparatus for processing image data associated with at least one input image, including a convolutional neural network, CNN, -based encoder configured to provide a plurality of hierarchical feature maps based on the image data, a decoder configured to provide output data based on the plurality of feature maps, wherein the decoder includes a convolutional long short-term memory, Conv-LSTM, module configured to sequentially process at least some of the plurality of hierarchical feature maps.

Abstract: In an exemplary embodiment, the present invention is an apparatus which includes at least one processor, and at least one memory including computer program code. The at least one memory and the computer program code are configured, with the at least one processor, to cause the apparatus to perform obtaining an image for segmentation; forming a trimap of at least a portion of the image, the trimap comprising a three-region map having a definite foreground, a definite background, and an unknown region separating the foreground and the background, said unknown region initially having a preselected width; in a plurality of iterations, changing the width of the unknown region; for each of the plurality of iterations, determining a quality measure; and selecting the width for the unknown region having the best quality measure. The apparatus may also include at least one camera for obtaining the image for segmentation.

Abstract: An apparatus configured to, in respect of a video provided to a user and wherein a plurality of comments are displayed such that they scroll across the video, based on the comments and one or more comment filtering rules, provide for comment filtering as each of the plurality of comments scrolls to meet a filter line, wherein the filter line extends across at least part of the video to define a first area overlaying the video and a non-overlapping second area overlaying the video, such that in the first area the plurality of comments are provided for display scrolling therethrough towards the filter line prior to comment filtering and in the second area the plurality of comments are provided for display with any comments of the plurality of comments that satisfy the one or more comment filtering rules removed from display.

Abstract: Aspects described herein relate to various methods, systems and apparatuses that may improve the accuracy of object classifications and object boundary definitions for a semantic segmentation technique. For example, the semantic segmentation technique may be based on a hierarchy of two or more layers. The two or more layers may include neural networks that analyze image data at different resolution scales. Each layer of the hierarchy may determine object boundary features and object class features. Each layer of the hierarchy may share its object boundary features and/or its object class features with one or more other layers in the hierarchy. In turn, each of other layers of the hierarchy may determine its object boundary features and/or its object class features based on the shared features.

Abstract: Aspects described herein relate to various methods, systems and apparatuses that may improve the accuracy of object classifications and object boundary definitions for a semantic segmentation technique. For example, the semantic segmentation technique may be based on a hierarchy of two or more layers. The two or more layers may include neural networks that analyze image data at different resolution scales. Each layer of the hierarchy may determine object boundary features and object class features. Each layer of the hierarchy may share its object boundary features and/or its object class features with one or more other layers in the hierarchy. In turn, each of other layers of the hierarchy may determine its object boundary features and/or its object class features based on the shared features.

Abstract: A method, comprising: receiving a video comprising video frames as an input; generating set of object proposals from the video; generating object tracklets comprising regions appearing in consecutive frames of the video, said regions corresponding to object proposals with a predetermined level of confidence; constructing a graph comprising the object proposals and superpixels grouped from the pixels of the frame; calculating a first cost function for superpixel likelihoods in terms of the object proposals in an object tracklet; calculating a second cost function for object proposal likelihoods in terms of the superpixels in the frame; minimizing the first and the second cost function with each other; computing posterior probabilities of each superpixel; and assigning each superpixel with an object class with a maximum posterior probability to constitute semantic object segmentation.

Abstract: An apparatus configured to, in respect of first and second virtual reality content (207, 209) each configured to provide imagery for a respective first and second virtual reality space for viewing in virtual reality; and based on first-user-viewing-experience information defining an appearance of an object of interest (205) that appears in the first virtual reality content as viewed, in virtual reality, by a first user (201), and defining a time-variant point of view from which the first user (201) viewed the object of interest (205); providing for display to a second user (202), the second user provided with a virtual reality view of the second virtual reality content (203), of imagery of the object of interest (204) superimposed into the virtual reality space of the second virtual reality content (203) such that the second user (202) is able, while viewing the second virtual reality content, to witness the object of interest (204) as it was viewed by the first user (201).

Abstract: In an exemplary embodiment, the present invention is an apparatus which includes at least one processor, and at least one memory including computer program code. The at least one memory and the computer program code are configured, with the at least one processor, to cause the apparatus to perform obtaining an image for segmentation; forming a trimap of at least a portion of the image, the trimap comprising a three-region map having a definite foreground, a definite background, and an unknown region separating the foreground and the background, said unknown region initially having a preselected width; in a plurality of iterations, changing the width of the unknown region; for each of the plurality of iterations, determining a quality measure; and selecting the width for the unknown region having the best quality measure. The apparatus may also include at least one camera for obtaining the image for segmentation.

Abstract: A method comprising: enabling a first user to define a message for display to at least a second user in association with a first three-dimensional scene viewed by the first user and viewed by or viewable by the second user, wherein the message comprises user-defined message content for display and message metadata, not for display, defining first three-dimensional spatial information; and enabling rendering of the user-defined message content in a second three-dimensional scene viewed by the second user, wherein the user-defined message content moves, within the second three-dimensional scene, along a three-dimensional trajectory dependent upon the first three-dimensional spatial information and three-dimensional spatial information of the second user.

Abstract: A method comprising performing image re-projection on each of a plurality of first images of a scene, thereby to generate a plurality of re-projected second images of the scene, wherein each first image of the scene is captured by a respective camera of a first multi-directional image capture apparatus and each second image of the scene is associated with a respective virtual camera; processing the plurality of second images based on a previously generated virtual three dimensional model of the scene, thereby to generate respective positions of the virtual cameras associated with the second images; and determining a position of the first multi-directional image capture apparatus based on one or more of the generated positions of the virtual cameras.

Abstract: A NN is trained using a cost function that places constraints on weights in the NN. The constraints are based on key(s) and cluster center(s) of the weights. The training embeds a capability to produce signature(s) corresponding to the key(s). Information is output that corresponds to the trained NN for testing a NN to determine if the tested NN is or is not verifiable as the trained NN. A NN is tested using the key(s) to determine output signature(s). The output signature(s) are compared, using a metric, with other signature(s) that correspond to the key(s). Based on the comparison, it is determined whether the NN is or is not verified as a known NN with the embedded capability to produce specific signatures corresponding to the key(s). In response to the NN being determined to be verified as the known NN, the NN is reported as being verified.

Abstract: An apparatus configured to, in respect of first and second virtual reality content (207, 209) each configured to provide imagery for a respective first and second virtual reality space for viewing in virtual reality; and based on first-user-viewing-experience information defining an appearance of an object of interest (205) that appears in the first virtual reality content as viewed, in virtual reality, by a first user (201), and defining a time-variant point of view from which the first user (201) viewed the object of interest (205); providing for display to a second user (202), the second user provided with a virtual reality view of the second virtual reality content (203), of imagery of the object of interest (204) superimposed into the virtual reality space of the second virtual reality content (203) such that the second user (202) is able, while viewing the second virtual reality content, to witness the object of interest (204) as it was viewed by the first user (201).

Abstract: An apparatus configured to, in respect of virtual reality content comprising video imagery configured to provide a virtual reality space wherein a virtual reality view presented to a user provides for viewing of the VR space; based on a comment made by and a virtual location of a commenting-user in the virtual reality space when the comment was made; provide for determination of a point of interest in the virtual reality space, the point of interest identified based on, at least, the virtual location of the commenting-user when the comment was made and semantic analysis of the comment to identify the point of interest surrounding the virtual location to which the comment refers, the point of interest associated with the comment thereby enabling the comment to be overlaid over the virtual reality view of the video imagery.

Abstract: A method comprising: enabling a first user to define a message for display to at least a second user in association with a first three-dimensional scene viewed by the first user and viewed by or viewable by the second user, wherein the message comprises user-defined message content for display and message metadata, not for display, defining first three-dimensional spatial information; and enabling rendering of the user-defined message content in a second three-dimensional scene viewed by the second user, wherein the user-defined message content moves, within the second three-dimensional scene, along a three-dimensional trajectory dependent upon the first three-dimensional spatial information and three-dimensional spatial information of the second user.

Abstract: An apparatus configured to, in respect of a video provided to a user and wherein a plurality of comments are displayed such that they scroll across the video, based on the comments and one or more comment filtering rules, provide for comment filtering as each of the plurality of comments scrolls to meet a filter line, wherein the filter line extends across at least part of the video to define a first area overlaying the video and a non-overlapping second area overlaying the video, such that in the first area the plurality of comments are provided for display scrolling therethrough towards the filter line prior to comment filtering and in the second area the plurality of comments are provided for display with any comments of the plurality of comments that satisfy the one or more comment filtering rules removed from display.

Abstract: The invention relates to a method and an apparatus implementing the method. The method comprises extracting region proposals from a media content; selecting a set of region proposals corresponding to an object in the media content; identifying objects of interest; determining an object-specific representation by an iterative tracking method; sampling positive examples from the set of tracked region proposal groups obtained from the iterative tracking method; and performing object segmentation.

Abstract: A method, comprising: receiving a video comprising video frames as an input; generating set of object proposals from the video; generating object tracklets comprising regions appearing in consecutive frames of the video, said regions corresponding to object proposals with a predetermined level of confidence; constructing a graph comprising the object proposals and superpixels grouped from the pixels of the frame; calculating a first cost function for superpixel likelihoods in terms of the object proposals in an object tracklet; calculating a second cost function for object proposal likelihoods in terms of the superpixels in the frame; minimizing the first and the second cost function with each other; computing posterior probabilities of each superpixel; and assigning each superpixel with an object class with a maximum posterior probability to constitute semantic object segmentation.