Netflix Inc.

Cisco Technology, Inc.

Intel Corporation

Zhi Li

Christos G Bampis

Joshua B Gahm

Anne Aaron

Anush Moorthy

Rong Pan

Xiaoqing Zhu

Te-Yuan Huang

Li-Heng Chen

David R Oran

Hao Hu

Aliasghar Eftekhar

Wei Qian

Andrew Berglund

Mengyuan Huang

Lavanya Sharan

Chaitanya Ekanadham

Martin Tingley

Julie Novak

John Heck

Ali C Begen

Bruce P Osler

Olufemi I Dosunmu

Mark Isenberger

Aditya Mavlankar

Harel Frish

Liwei Guo

Todd Goodall

Jonathan Doylend

Reece DeFrees

Aimin Ding

Kelly Christopher Magruder

Seth M Slavin

Josh Gahm

Srinivasan Sethuraman

Gregory Lovell

Sungbong Park

Cung Tran

Andrew J Berglund

In various embodiments an endpoint application reconstructs downscaled videos. The endpoint application accesses metadata associated with a portion of a downscaled video that has a first resolution and was generated using a trained downscaling convolutional neural network (CNN). The endpoint application determines, based on the metadata, an upscaler that should be used when upscaling the portion of the downscaled video. The endpoint application executes the upscaler on the portion of the downscaled video to generate a portion of a reconstructed video that is accessible for playback and has a second resolution that is greater than the first resolution.

Techniques for reconstructing downscaled video content

In various embodiments, a training application generates a trained perceptual quality model that estimates perceived video quality for reconstructed video. The training application computes a pixels-per-degree value based on a normalized viewing distance and a display resolution. The training application computes a set of feature values corresponding to a set of visual quality metrics based on a reconstructed video sequence, a source video sequence, and the pixels-per-degree value. The training application executes a machine learning algorithm on the first set of feature values to generate the trained perceptual quality model. The trained perceptual quality model maps a particular set of feature values corresponding to the set of visual quality metrics to a particular perceptual quality score.

Techniques for generating a perceptual quality model for predicting video quality across different viewing parameters

Embodiments described herein may be related to apparatuses, processes, and techniques directed to a planar germanium photodetector that includes n-type and p-type amorphous silicon deposits on a germanium slab. During operation, a uniform electrical field is formed across the germanium bulk between the amorphous silicon deposits. Other embodiments may be described and/or claimed.

Planar germanium photodetector

In various embodiments, a tunable VMAF application reduces an amount of influence that image enhancement operations have on perceptual video quality estimates. In operation, the tunable VMAF application computes a first value for a first visual quality metric based on reconstructed video content and a first enhancement gain limit. The tunable VMAF application computes a second value for a second visual quality metric based on the reconstructed video content and a second enhancement gain limit. Subsequently, the tunable VMAF application generates a feature value vector based on the first value for the first visual quality metric and the second value for the second visual quality metric. The tunable VMAF application executes a VMAF model based on the feature value vector to generate a tuned VMAF score that accounts, at least in part, for at least one image enhancement operation used to generate the reconstructed video content.

Techniques for limiting the influence of image enhancement operations on perceptual video quality estimations

A balanced photodetector may include: a balanced photodetector including a first photodiode and a second photodiode coupled with one another at a common node, wherein the first photodiode has a first effective responsivity and the second photodiode has as second effective responsivity; and a control circuit configured to set an operating parameter of the balanced photodetector to compensate for a difference between the first effective responsivity and the second effective responsivity.

Balanced photodetector and methods thereof

In one embodiment, an apparatus includes a substrate, an oxide layer on the substrate, a silicon layer on the oxide layer, which includes a waveguide region and etched regions adjacent to the waveguide region, a germanium layer on the silicon layer and adjacent the waveguide region of the silicon layer, and a resistive element adjacent to the germanium layer to provide heat to the germanium layer in response to a current applied to the resistive element.

Germanium waveguide photodetector with integrated heater

In various embodiments, a quality inference application estimates perceived video quality for reconstructed video. The quality inference application computes a set of feature values corresponding to a set of visual quality metrics based on a reconstructed frame, a source frame, a display resolution, and a normalized viewing distance. The quality inference application executes a trained perceptual quality model on the set of feature values to generate a perceptual quality score that indicates a perceived visual quality level for the reconstructed frame. The quality inference application performs one or more operations associated with an encoding process based on the perceptual quality score.

Techniques for predicting video quality across different viewing parameters

In various embodiments, a bootstrapping training subsystem performs sampling operation(s) on a training database that includes subjective scores to generate resampled dataset. For each resampled dataset, the bootstrapping training subsystem performs machine learning operation(s) to generate a different bootstrap perceptual quality model. The bootstrapping training subsystem then uses the bootstrap perceptual quality models to quantify the accuracy of a perceptual quality score generated by a baseline perceptual quality model for a portion of encoded video content. Advantageously, relative to prior art solutions in which the accuracy of a perceptual quality score is unknown, the bootstrap perceptual quality models enable developers and software applications to draw more valid conclusions and/or more reliably optimize encoding operations based on the perceptual quality score.

Quantifying perceptual quality model uncertainty via bootstrapping

Embedded three-dimensional electrode capacitors, and methods of fabricating three-dimensional electrode capacitors, are described. In an example, an integrated circuit structure includes a first metallization layer above a substrate, the first metallization layer having a first conductive structure in a first dielectric layer, the first conductive structure having a honeycomb pattern. An insulator structure is on the first conductive structure of the first metallization layer. A second metallization layer is above the first metallization layer, the second metallization layer having a second conductive structure in a second dielectric layer, the second conductive structure on the insulator structure, and the second conductive structure having the honeycomb pattern.

Embedded three-dimensional electrode capacitor

In various embodiments, a training application trains a convolutional neural network to downsample images in a video encoding pipeline. The convolution neural network includes at least two residual blocks and is associated with a downsampling factor. The training application executes the convolutional neural network on a source image to generate a downsampled image. The training application then executes an upsampling algorithm on the downsampled image to generate a reconstructed image having the same resolution as the source image. The training application computes a reconstruction error based on the reconstructed image and the source image. The training application updates at least one parameter of the convolutional neural network based on the reconstruction error to generate a trained convolutional neural network. Advantageously, the trained convolution neural network can be implemented in a video encoding pipeline to mitigate visual quality reductions typically experienced with conventional video encoding pipelines that implement conventional downsampling techniques.

Machine learning techniques for video downsampling

In various embodiments, a hindsight application computes a hindsight metric value for evaluation of a video rate selection algorithm. The hindsight application determines a first encoding option associated with a source chunk of a media title based on a network throughput trace and a buffer trellis. The hindsight application determines that the first encoding option is associated with a buffered duration range. The buffered duration range is also associated with a second encoding option that is stored in the buffer trellis. After determining that the first encoding option is associated with a higher visual quality than the second encoding option, the hindsight application stores the first encoding option instead of the second encoding option in the buffer trellis to generate a modified buffer trellis. Finally, the hindsight application computes a hindsight metric value associated with a sequence of encoded chunks of the media title based on the modified buffer trellis.

Techniques for determining an upper bound on visual quality over a completed streaming session

In various embodiments, a perceptual quality application determines an absolute quality score for encoded video content viewed on a target viewing device. In operation, the perceptual quality application determines a baseline absolute quality score for the encoded video content viewed on a baseline viewing device. Subsequently, the perceptual quality application determines that a target value for a type of the target viewing device does not match a base value for the type of the baseline viewing device. The perceptual quality application computes an absolute quality score for the encoded video content viewed on the target viewing device based on the baseline absolute quality score and the target value. Because the absolute quality score is independent of the viewing device, the absolute quality score accurately reflects the perceived quality of a wide range of encoded video content when decoded and viewed on a viewing device.

Device-consistent techniques for predicting absolute perceptual video quality

In various embodiments, an encoding ladder application generates encoding ladders for encoding media titles. In operation, the encoding ladder application generates a first convex hull representing encoding tradeoffs between quality and bitrate when encoding a media title at a first resolution; The encoding ladder application generates a second convex hull representing encoding tradeoffs between quality and bitrate when encoding the media title at a second resolution. Based on the first convex hull and the second convex hull, the encoding ladder application generates an overall convex hull. Subsequently, the encoding ladder application generates an encoding ladder for the media title based on at least the overall convex hull and a ladder requirement. Advantageously, the tradeoffs between quality and bitrate represented by the encoding ladder are customized for the media title. Consequently, encoding inefficiencies attributable to conventional fixed-bitrate ladders can be reduced.

Techniques for generating per-title encoding ladders

In various embodiments, a prediction application computes a quality score for re-constructed visual content that is derived from visual content. The prediction application generates a frame difference matrix based on two frames included in the re-constructed video content. The prediction application then generates a first entropy matrix based on the frame difference matrix and a first scale. Subsequently, the prediction application computes a first value for a first temporal feature based on the first entropy matrix and a second entropy matrix associated with both the visual content and the first scale. The prediction application computes a quality score for the re-constructed video content based on the first value, a second value for a second temporal feature associated with a second scale, and a machine learning model that is trained using subjective quality scores. The quality score indicates a level of visual quality associated with streamed video content.

Techniques for modeling temporal distortions when predicting perceptual video quality

Growing community of inventors

Mountain View, CA, United States of America

Zhi Li