Microsoft Technology Licensing, LLC

Daniel Gilbert Kennett

Jonathan Ross Hoof

Anis Ahmad

Beibei Cheng

Martin Jon Irwin Fuller

Benjamin William Walker

Matthew Gordon

Andrew C Flavell

Eric Cameron Walston

Ben Walker

Mark Edward Rabold

Michael Anthony Hall

Kameron Wade Rausch

Abdelrehim Ahmed

David C McCarthy

A silhouette-based limb finder may be used to detect limbs from a camera image. This limb determination may be used to control an application, such as a game, or a combination with other image processing. A first distance field indicating a distance from the edge of a silhouette in an image and a second distance field indicating distance from a location in the silhouette may be used to generate a path from an extremity point on the silhouette to the location. This path then may be used to determine a limb in the silhouette. This allows tracking of limbs even for hard to detect player poses.

Silhouette-based limb finder determination

The present disclosure relates to systems, methods, and computer-readable media for identifying a variety of events that occur within a gaming session and generating event reports based on the identified events. For example, a gaming service (e.g., a cloud gaming server) can leverage content analysis and event recognizer services on a cloud computing system to detect one or more in-game events based on gaming content (e.g., video content, audio content, controller inputs) that is delivered to a client system. Systems described herein can train and implement event recognizers trained to track various in-game events across multiple gaming applications. Based on the tracked events, the systems described herein can generate event reports for events, individual users, and groups of users of the cloud computing system.

Tracking in-game events and generating event reports across multiple gaming applications

The present disclosure relates to systems, methods, and computer-readable media to selectively enhance digital image and video content. For example, systems disclosed herein can encode original video content to compress and decompress the original video content. Systems described herein can further identify area of interest information for use in identifying portions of decompressed video content to analyze and remove one or more compression artifacts found therein. Systems described herein can further enhance the decompressed video content by increasing resolution for display. By identifying areas of interest and selectively enhancing digital video content, the systems described herein can reduce consumption of bandwidth and processing resources while maintaining high visual quality of the digital content.

Selectively enhancing compressed digital content

The present disclosure relates to systems, methods, and computer-readable media for identifying a variety of events that occur within a gaming session and implementing one or more in-game enhancements based on the identified events. For example, a gaming service (e.g., a cloud gaming server) can leverage content analysis and event recognition services on a cloud computing system to detect one or more in-game events based on gaming content (e.g., video content, audio content, controller inputs) that is delivered to a client system. Systems described herein can cause in-game enhancements to be implemented on a client system in response to detected events. Detection of events and implementation of the various enhancements may be performed without significant redevelopment of gaming applications, even where a gaming application has been previously shipped.

Enhancing gaming content for previously developed gaming applications

Methods and devices for generating hardware compatible compressed textures may include accessing, at runtime of an application program, graphics hardware incompatible compressed textures in a format incompatible with a graphics processing unit (GPU). The methods and devices may include converting the graphics hardware incompatible compressed textures directly into hardware compatible compressed textures usable by the GPU using a trained machine learning model.

Machine learning applied to textures compression or upscaling

Techniques for dynamic exercise content are described. In implementations, exercise content is provided that includes a variety of different selectable exercise segments that can be individually selected and played back to generate an exercise routine. For example, particular exercise segments can be selected based on user-specified exercise goals, the physical abilities of a particular user, based on various types of feedback, and so on. To assist in the selection of particular exercise segments, exercise segments can be individually tagged with descriptive information, such as using metadata tags. Embodiments can also provide a variety of different types of performance-related feedback to a user during an exercise routine.

Dynamic exercise content

Methods and devices for real time texture compression may include accessing graphics hardware incompatible compressed textures in a format incompatible with the GPU, and a metadata file associated with the graphics hardware incompatible compressed textures, wherein the metadata file includes at least one hint that provides information to use for compression of decompressed textures from the graphics hardware incompatible compressed textures into hardware compatible compressed textures. The methods and devices may include converting the graphics hardware incompatible compressed textures into the decompressed textures. The methods and devices may include selectively compressing the decompressed textures into the hardware compatible compressed textures usable by the GPU according to the at least one hint from the metadata file. The methods and devices may include transmitting the hardware compatible compressed textures to the GPU.

Reducing the search space for real time texture compression

A facial skin mask may be generated based on isolating a head part in a captured image, removing a first pixel that is indicative of non-skin from the head part in the captured image, and removing a second pixel that is indicative of having a high velocity from the head part in the captured image. Heart rate may be detected based on the change of color of the pixels of the generated facial skin mask.

Facial skin mask generation for heart rate detection

Heartrate tracking is performed entirely optically without the subject being required to wear any monitoring equipment by processing a combination of signals representing frames of video of the sinusoidal motion of a subject's facial skin color changes captured by both IR and visible light (e.g., RGB&#x2014;red/green/blue) cameras. The IR and RGB graphs that result from the processing are perfectly phase-shifted so that when the IR signal is going down in amplitude, the RGB signal is going up. Such phase-shifting enables the optical heartrate tracking to utilize diverse input feeds so that a tracked signal is accepted as the user's true heartrate when both IR and RGB signals are well correlated.

Optical heartrate tracking

Embodiments described herein use depth images to extract user behavior, wherein each depth image specifies that a plurality of pixels correspond to a user. In certain embodiments, one or more average extremity positions of a user, which can also be referred to as average positions of extremity blobs, are extracted from a depth image. An application is then updated based on the average positions of extremity blobs.

Growing community of inventors

Bellevue, WA, United States of America

Daniel Gilbert Kennett