Abstract: Restoration methods and systems are disclosed for video remastering. Techniques disclosed include receiving a video sequence. For each frame of the video sequence, techniques disclosed include encoding, by a degradation encoder, a video content associated with the frame into a latent vector. The latent vector is a representation of the degradation present in the video content; the degradation present in the video content includes one or more degradation types. Based on the latent vector and the video content, techniques disclosed further include generating, by a backbone network, one or more feature maps, and, then, restoring the frame based on the one or more feature maps.

Abstract: A system for performing synergistic object tracking and pattern recognition for event representation includes a computing platform having processing hardware and a system memory storing a software code. The processing hardware is configured to execute the software code to receive event data corresponding to one or more propertie(s) of an object, to generate, using the event data, a location data estimating a location of each of multiple predetermined landmarks of the object, and to predict, using one or both of the event data and the location data, a pattern corresponding to the propertie(s) of the object. The processing hardware is further configured to execute the software code to update, using the predicted pattern, the location data, and to merge the updated location data and the predicted pattern to provide merged data.

Abstract: Techniques are disclosed for creating digital faces. In some examples, an anatomical face model is generated from a data set including captured facial geometries of different individuals and associated bone geometries. A model generator segments each of the captured facial geometries into patches, compresses the segmented geometry associated with each patch to determine local deformation subspaces of the anatomical face model, and determines corresponding compressed anatomical subspaces of the anatomical face model. A sculpting application determines, based on sculpting input from a user, constraints for an optimization to determine parameter values associated with the anatomical face model. The parameter values can be used, along with the anatomical face model, to generate facial geometry that reflects the sculpting input.

Abstract: Techniques for machine learning optimization are provided. A video comprising a plurality of segments is received, and a first segment of the plurality of segments is processed with a machine learning (ML) model to generate a plurality of tags, where each of the plurality of tags indicates presence of an element in the first segment. A respective accuracy value is determined for each respective tag of the plurality of tags, where the respective accuracy value is based at least in part on a maturity score for the ML model. The first segment is classified as accurate, based on determining that an aggregate accuracy of tags corresponding to the first segment exceeds a predefined threshold. Upon classifying the first segment as accurate, the first segment is bypassed during a review process.

Abstract: Various embodiments set forth systems and techniques for training a landmark model. The techniques include determining, using the landmark model, a first landmark in a set of first landmarks associated with a first image; performing, on the first image, a first perturbation to obtain a second image; determining, using the landmark model, a second landmark in a set of second landmarks associated with the second image; determining, based on a first distance between the first landmark and the second landmark, a first loss function; and updating, based on the first loss function, a first parameter of the landmark model.

Abstract: There is provided a method for use by a playback system to play back encrypted content associated with a first globally unique identifier (ID). The method includes storing an encrypted content in a memory being under control of a memory controller having a second globally unique ID, storing the first globally unique ID, transmitting a request from the memory controller to a license server, the request comprising the first globally unique ID, the second globally unique ID and a request for a content license to play back the encrypted content, receiving the content license enabling a right for the playback of the encrypted content, the content license including content keys and being encrypted to be decryptable according to a memory controller secret associated with the second globally unique ID, decrypting the content license, according to the memory controller secret to recover the content keys to a playback module.

Abstract: A virtual puppeteering system includes a portable device including a camera, a display, a hardware processor, and a system memory storing an object animation software code. The hardware processor is configured to execute the object animation software code to, using the camera, generate an image in response to receiving an activation input, using the display, display the image, and receive a selection input selecting an object shown in the image. The hardware processor is further configured to execute the object animation software code to determine a distance separating the selected object from the portable device, receive an animation input, identify, based on the selected object and the received animation input, a movement for animating the selected object, generate an animation of the selected object using the determined distance and the identified movement, and render the animation of the selected object.

Abstract: Methods for improving operation of a user device executing an application. The methods include collecting a first set of data corresponding to a run time environment of the application, collecting a second set of data corresponding to a crash of the application, identifying a cause of the crash based on the first set of data and a second set of data and determining the cause of the crash is associated with an application feature corresponding to a feature flag.

Abstract: Techniques for buffering data over high bandwidth networks are provided. A first portion of data is downloaded into a buffer at a first download speed via a first network. The first portion of data is outputted from the buffer. Upon determining that data is being downloaded, via a second network, at a second download speed greater than the first download speed, a size of the buffer is expanded, and a second portion of data is downloaded into the expanded buffer.

Abstract: The embodiments herein use a corrector plate or a light field display in an AR/VR display device to compensate for sub-optimal collimation at the edge of the FOV. In one embodiment, the corrector plate is disposed between the collimator and the viewer so that the light at the edge of the FOV can be corrected so that the aberrations mentioned above do not occur. In another embodiment, rather than using a corrector plate, the AR/VR display device can include a light field display that can use color intensity to pre-distort emitted light to compensate for sub-optimal collimation at the edge of the FOV. In this manner, the AR/VR display device can mitigate aberrations or distortions as the user moves her eyes relative to the display device.

Abstract: According to one implementation, a method of guiding an interaction between a companion module and a user includes identifying a media content for playout by a media player device, transmitting, to a remote server, an interaction schema request identifying the media content and the companion module, and receiving, from the remote server, behavioral manifest data including an instruction for guiding the interaction between the companion module and the user. Such a method also includes obtaining a play head state of the media player device and a playout timecode state of the media content, and identifying a user interaction behavior for the companion module based on the behavioral manifest data, the play head state of the media player device, and the playout timecode state of the media content.

Abstract: This disclosure presents systems and methods to improve user brushing behavior. Exemplary implementations may: effectuate presentation of audio content via an audio output device, the audio content being characterized by an audio profile; obtain output signals from one or more sensors coupled to a user device, the output signals conveying user input information, the user input information defining user input by a user via the user device; determine, based on the user input information, a user profile, the user profile including one or more user values of the one or more audio attribute; compare the user profile with the audio profile to determine, for individual audio attributes, one or more differences; determine one or more scores reflecting the one or more differences; and/or perform other operations.

Abstract: A method of generating a training data set for training an image matting machine learning model includes receiving a plurality of foreground images, generating a plurality of composited foreground images by compositing randomly selected foreground images from the plurality of foreground images, and generating a plurality of training images by compositing each composited foreground image with a randomly selected background image. The training data set includes the plurality of training images.

Abstract: Systems and methods for providing machine-learning enabled user-specific evaluations are disclosed. Implementations include obtaining a first set of evaluation data from a user interface, obtaining a first set of target-descriptive data including target-specific characteristics objectively describing the evaluation targets, and training, with a machine-learning algorithm, a user-specific evaluation profile indicating evaluation patterns relative to the first set of evaluation data and the first set of target-specific characteristics. Implementations include applying the user-specific evaluation profile to a second set of target-descriptive data to predict a user-specific evaluation.

Abstract: A modular architecture is provided for denoising Monte Carlo renderings using neural networks. The temporal approach extracts and combines feature representations from neighboring frames rather than building a temporal context using recurrent connections. A multiscale architecture includes separate single-frame or temporal denoising modules for individual scales, and one or more scale compositor neural networks configured to adaptively blend individual scales. An error-predicting module is configured to produce adaptive sampling maps for a renderer to achieve more uniform residual noise distribution. An asymmetric loss function may be used for training the neural networks, which can provide control over the variance-bias trade-off during denoising.

Abstract: A system and method for model-based camera tracking and image occlusion removal for a camera viewing a sports field (or other scene) includes receiving a synthesized data set comprising at least one empty field image of the field, the empty field image with at least one occlusion graphic, and camera parameters corresponding to the empty field image, training a neural network model to estimate the empty field image and the corresponding camera parameters by providing the model with an input training image comprising the empty field image with occlusion graphic, and providing the model with model output targets comprising the empty field image and the corresponding camera parameters as targets for the model, receiving by the neural network model, alive input image comprising a view of the field with live occlusions, and providing by the neural network model, using trained model parameters, estimated live camera parameters or an estimated empty field image associated with the live input image.

Abstract: A method to project content from a moving projector. The method includes analyzing an object to identify a projection surface; determining a first position of a projector relative to the projection surface; modifying a first frame of a content for projection onto the projection surface based on the projection surface and the first position of the projector; projecting the first frame of the content from the projector onto the projection surface; determining a second position of the projector relative to the projection surface; modifying a second frame of the content based on the projection surface and the second position; and projecting the second frame of the content from the projector onto the projection surface.

Abstract: A method includes displaying, on a touchscreen, a video comprising a video frame and determining, based on a saliency map of the video frame, a region of interest in the video frame. The method also includes detecting a touch on a region of the touchscreen while the video frame is displayed and generating a haptic response in response to determining that the region of the touchscreen overlaps with the region of interest.

Abstract: A method includes storing a plurality of scores and a plurality of session identifiers for a game. The plurality of scores are generated by monitoring player activity during game sessions identified by the plurality of session identifiers and by applying rules of the game to determine scoring events based on the player activity. The method also includes generating, during a game session, a code for a player, providing the code to the player before the player leaves the game, and receiving, from a device, a target session identifier generated by the device based on the code and in response to receiving the target session identifier. The method further includes retrieving a score of the plurality of scores generated during a game session identified by a session identifier of the plurality of session identifiers that matches the target session identifier and communicating, to the device, the retrieved score.