41 results about "Streaming application" patented technology

In various examples, techniques for optimizing object detection models are described herein. Systems and methods are disclosed that process sensor data using a backbone of a machine learning model(s) in order to generate feature maps at different resolutions. The systems and methods then use the machine learning model(s) to generate a vector based at least in part on one or more of the feature maps. For example, if the backbone generates four feature maps, then the machine learning model(s) may generate the vector using two feature maps from the four feature maps. The systems and methods then process the vector using a transformer of the machine learning model(s) in order to generate data representing a class label(s) for an object(s) depicted by an image represented by the sensor data and / or a location(s) of the object(s) within the image.

An apparatus and method for efficiently performing efficient data storage and data transfer of machine learning data. In various implementations, a host processing circuit of a computing system executes a machine learning (ML) application. The application includes a computational graph that indicates the computational order of the ML nodes, layers, and stages of the ML model. The host processing circuit translates function calls in the application to commands particular to an accelerator circuit. The accelerator circuit preloads weights to be used by ML nodes of the ML model by retrieving compressed weights from a storage device different from system memory. The accelerator circuit uses a streaming application programming interface (API) and bypasses the host processing circuit to retrieve the compressed weights from the storage device. The accelerator circuit decompresses the retrieved weights and executes the ML node using the decompressed weights.

This application relates to the fields of cloud gaming and remote streaming display, and in particular to a boot-up method, apparatus, device, and medium. The method includes: in the initramfs stage, a startup animation program determines the optimal display mode, which can be uniformly displayed in subsequent display processes. Furthermore, during the control switching process in each stage, the current frame of the buffer of the previous program is continuously output and used as the first output buffer frame of the subsequent program, ensuring screen consistency and achieving smooth connection of multi-stage display. An early display proxy process and a parallel initialization mechanism are introduced to achieve continuous visual output across the entire link from system power-on to the display of the first frame of the target streaming application, effectively avoiding black screen, flickering, and screen jump phenomena during startup, and significantly improving the user's perceived experience of system startup speed and visual quality.

According to an aspect, a method includes rendering a user interface on an extended reality device, the user interface identifying a media item for playback using a streaming application, in response to selection of the media item for playback, initiating a display of immersive imagery related to the media item on the extended reality device, and transmitting a request to the streaming application. The request includes at least one parameter about the immersive imagery that causes the streaming application to render a display panel within the immersive imagery using the at least one parameter, the display panel displaying content of the media item.

Disclosed are apparatuses, systems, and techniques for implementing fractionalized data transfers between processing devices in real-time data generating and streaming applications. The techniques include but are not limited to processing, by a first processing device, image data to generate a plurality of portions of an image, responsive to generating a first portion of the plurality of portions of the image, storing the first portion in a first memory device of the first processing device, setting a completion indicator for the first portion, and causing the first portion to be provided to a second processing device.

There may be provided a method of streaming multimedia content encoded using an enhancement coding scheme and a method of packaging multimedia content encoded using an enhancement coding scheme for streaming, comprising retrieving an constructing a manifest file, the manifest file comprising metadata for constructing a locator to segments of the encoded multimedia content, the metadata comprising: a first representation pointing to a first version of the multimedia content encoded using a base coding scheme; and, location metadata for a first enhancement layer of the encoded multimedia content associated with the first version. A client device, packager, system and computer readable medium are also provided.

Disclosed are apparatuses, systems, and techniques that implementing fractionalized data transfers between processing devices in real-time data generating and streaming applications. The techniques include but are not limited to processing, by a first processing device, an image data to generate a plurality of portions of an image, responsive to generating a first portion of the plurality of portions of the image, storing the first portion in a first memory device of the first processing device, setting a completion indicator for the first portion, and causing the first portion to be provided to a second processing device.

A method of controlling energy consumed by a multimedia streaming application executed in a multimedia streaming system, the method comprising: receiving (358), at an end-user device, a message including information representative of energy consumed by at least one node of the multimedia streaming system when applying at least one process on content having a first profile delivered by the multimedia streaming application prior to delivering the content having the first profile to the end-user device; and requesting (359), at the end-user device, content having a second profile based on the received information representative of the energy consumed by the at least one node.

Techniques described herein are related to updating streaming application software. For example, data from the streaming application software are transformed into a format for an engine to generate one or more results. The one or more results may comprise information related to the data from the streaming application software.

A method and system for generating test cases for a streaming application is disclosed. An execution plan for testing a functionality of the streaming application is received. A machine-learning module determines a workflow data associated with the functionality using at least one of a user profile library, an intent library, a device library, an input devices library, and a navigation library. A first and a second portion of the workflow are categorized into converging patterns and diverging patterns based on consistencies and variations in the workflow respectively. A test components generation module generates a first set and a second set of test components from the converging and the diverging patterns. A unique test cases corresponding to each of the plurality of platforms, operating systems, end devices, user profiles, and versions of the streaming application is then generated at runtime.

This invention provides a program, an information processing device, an information processing method, and an information processing system that simplify the editing operations of streamed videos and improve the usability of streaming applications. [Solution] A program to be executed by an information processing device for distributing a video in which images of performers are placed in a background image or a virtual space background includes: a storage process for storing a plurality of object candidates, which are candidates for objects that constitute the video, and a plurality of background candidates, which are candidates for backgrounds; a configuration process for composing each of the plurality of background candidates using at least a portion of the plurality of object candidates; and a setting process for setting the selected background candidate as the background in the video in response to an operation by the video distributor to select one of the plurality of background candidates.

A method and system for generating test cases for a streaming application is disclosed. An execution plan for testing a functionality of the streaming application is received. A machine-learning modul

Systems and methods herein address reference frame selection in video streaming applications using one or more processing units to replace, during receipt of an encoded video stream, a first set of frames stored in a cache with a second set of frames based at least in part on an indication within the encoded video stream that the second set of frames includes a non-blurred frame (NBF).

An apparatus and method for efficiently performing efficient data storage and data transfer of machine learning data. In various implementations, a host processing circuit of a computing system executes a machine learning (ML) application. The application includes a computational graph that indicates the computational order of the ML nodes, layers, and stages of the ML model. The host processing circuit translates function calls in the application to commands particular to an accelerator circuit. The accelerator circuit preloads weights to be used by ML nodes of the ML model by retrieving compressed weights from a storage device different from system memory. The accelerator circuit uses a streaming application programming interface (API) and bypasses the host processing circuit to retrieve the compressed weights from the storage device. The accelerator circuit decompresses the retrieved weights and executes the ML node using the decompressed weights.

Devices, systems, and methods for machine learning-based micro-category generation may include a method including identifying, for a respective cluster of content titles or items, a top-K most representative group of the content titles or items; generating, by a machine learning model, based on a first input prompt, the common theme of the top-K most representative group; generating, by the machine learning model, based on the second input prompt and filtering, a group of content titles or items matching the common theme; generating, by the machine learning model, based on a third input prompt, a name for the group of content titles or items matching the common theme; and presenting the group of content titles or items matching the common theme and the name for the group of content titles or items matching the common theme via a user interface of a streaming media application.

Technologies directed to Internet Protocol (IP) tunnels for cloud-based applications on a streaming device are described. One method receives a request to execute a streaming application hosted by the remote computing device. The method establishes a first IP tunnel with the remote computing device using a first virtual interface. The method establishes a second IP tunnel with a container executing the streaming application using a second virtual interface. The method establishing routing between the second virtual interface and a local network interface. The method communicates data with the streaming application over the second IP tunnel.

Apparatuses, systems, and techniques to generate and / or perform one or more software containers. In at least one embodiment, one or more software containers configure communication, such as to perform streaming applications. In at least one embodiment, a processor comprising one or more circuits generates one or more software containers, wherein one or more software containers are to automatically provide network configuration information to one or more proxy servers.

The present disclosure relates to entropy-based pre-filtering using neural networks for streaming applications. In various examples, a deep neural network (DNN)-based pre-filter for content streaming applications is used to dynamically adjust scene entropy (e.g., complexity) in response to changing network or system conditions for end-user devices. For example, where network and / or system performance issues or degradations are identified, the DNN can be implemented as a frame pre-filter to reduce the complexity or entropy of frames prior to streaming, allowing the frames to be streamed at a reduced bitrate without requiring changes to resolution. The DNN-based pre-filter can be adjusted to preserve image detail along object, boundary, and / or surface edges so that scene navigation - e.g., by a user participating in an application instance - can be easier, more natural for the user.

Systems and methods for logging OTT device requests and responses are provided. A method may include intercepting, via a logging system, one or more ad requests to be satisfied by an ad server. Data of the ad requests is logged, via the logging system. The request is forwarded to the ad server, which may respond with an ad server response that may also be logged by the logging system.

Embodiments include methods, electronic device, storage medium, and computer program for content creation in media streaming. In one embodiment, a method comprises: generating semantic information based on a first set of media frames, the semantic information to identify objects within the first set of media frames that are from a media stream of a media streaming application to be experienced by a user. The method continues with transmitting the semantic information to a server through a wireless network and receiving updated semantic information from the server through the wireless network, where the updated semantic information is generated to anticipate movement of the user and is generated based on the semantic information; and generating a second set of media frames based on the updated semantic information using a generative artificial intelligence model, the second set of media frames to be included in the media stream.

Systems and methods for logging OTT device requests and responses are provided. A method may include intercepting, via a logging system, one or more ad requests to be satisfied by an ad server. Data of the ad requests is logged, via the logging system. The request is forwarded to the ad server, which may respond with an ad server response that may also be logged by the logging system.

In various examples, processing content data using graphics processing units for video streaming systems and applications is described herein. Systems and methods are disclosed that offload at least a portion of the processing that is typically performed by a central processing unit (CPU) to a graphics processing unit (GPU). For example, and for a streaming application, the GPU may initially generate and then encode a content stream (e.g., a video stream, an audio stream, etc.). In some examples, the GPU generates and / or encodes the content stream based on input data received from one or more client devices. The GPU may then perform additional processing associated with the encoded content stream, such as packetization, forward error correction (FEC), encryption, and / or any other processing. The CPU may then use packet pacing when causing the processed content stream to be streamed to the client device(s).

To provide a neural network-based pre-filter for content streaming applications.SOLUTION: In various examples, a DNN-based pre-filter for content streaming applications is used to dynamically adapt scene entropy in response to changing network or system conditions of an end-user device. For example, where network and / or system performance issues or degradation are identified, the DNN may be implemented as a frame pre-filter to reduce the complexity or entropy of the frame prior to streaming, thereby allowing the frame to be streamed at a reduced bit rate without requiring a change in resolution. The DNN-based pre-filter may be tuned to maintain image detail along an object, or the like such that scene navigation may be easier and more natural to the user.SELECTED DRAWING: Figure 1A

This invention discloses a user behavior path analysis method, apparatus, device, and storage medium. The method includes: acquiring user operation tracking data and storing the operation tracking data in a data warehouse to obtain operation tracking logs; consuming and processing the operation tracking data through a Spark Streaming application, and associating it with user ID information, user branch information, and product dimension tables to obtain a real-time table; using a Spark scheduled task to acquire the operation tracking logs, and associating the operation tracking logs with user ID information, user branch information, and product dimension tables to obtain an offline table; and analyzing user behavior paths based on the real-time table and the offline table. Compared with existing technologies, this method not only improves the accuracy of user behavior path analysis but also effectively enhances the timeliness of data in user behavior path analysis.

A system and method of playing streaming audio by a wireless audio delivery device, when the audio delivery device is in a wireless limited environment, includes connecting with a streaming media service via a streaming media application executing on an electronic communication device. A system and method further include receiving, by the streaming media application, streaming audio from the streaming media service, and storing the streaming audio in a memory associated with the audio delivery device when the audio delivery device is not in the wireless limited environment. The system and method further includes playing, by the audio delivery device, the streaming audio stored in the memory.

Implementation of an extended coding scheme in streaming applications Methods for streaming multimedia content encoded using an extended coding scheme and methods for packaging multimedia content encoded using an extended coding scheme for streaming may be provided. These include obtaining and constructing a manifest file, the manifest file containing metadata for constructing locators to segments of the encoded multimedia content, the metadata including a first representation pointing to a first version of the multimedia content encoded using a base coding scheme, and location metadata for a first extension layer of the encoded multimedia content associated with the first version. Client devices, packagers, systems, and computer-readable media are also provided.

Methods, device and data stream are provided to generate, transmit and decode haptic data. The present principles generally relate to the domain of encoding haptic data. In particular, the present principles relate to a block-based structure for haptic data allowing to process chunks of data in parallel for scalable network distribution like streaming applications.

The embodiment of the invention provides a video storage method, device, system and equipment for a camera and a medium, and relates to the technical field of data transmission, the method is applied to a management engine of a streaming application system, and the method comprises the following steps: in response to a received alarm event, determining target time information based on target occurrence time of the alarm event; detecting whether a flow pulling task meeting a preset merging condition exists or not currently; if yes, time information of the pull flow task is updated based on target time information, so that the pull flow task of which the corresponding alarm event contains the alarm event is obtained; if not, generating a pull flow task corresponding to the alarm event based on the target time information; and in response to the task execution time for reaching any stream pulling task, based on the time information of the stream pulling task, performing stream extraction from the camera through the stream extraction application to obtain an event video of the alarm event corresponding to the stream pulling task. According to the invention, the event video of the alarm event can be stored to the central management equipment end.

The disclosure relates generally to methods and systems to synergize context of end-user with quality-of-experience (QoE) of live video feed. Conventional techniques mostly lack synergy between encoder / decoder and underlying protocol, and they mostly employ motion vector-based encoding, thereby making them not so suitable for latency critical application. In most of the telepresence / teleoperation based live video streaming applications, at a time user's attention is focused on certain region of streamed video and not the entire frame. But most of existing techniques under challenges network conditions, undermine quality evenly throughout frame without giving importance to end-user's instantaneous region of interest. Existing mechanisms that try to alleviate this problem using foveated rendering are computationally expensive hence cannot be deployed easily on real life robotics platforms. The present disclosure attempts to alleviate all above challenges through end-user foveation centric spatio-temporal bitrate adaptation scheme tightly entangled with underlying protocol to achieve dynamic foveated rendering.

The present disclosure relates to a streaming application environment with recovery of lost or delayed input events. In a streaming application environment coupled to a remote device over a packet-switched network, a corrective event can be synthesized from input events received from the remote device and injected into the virtual streaming application to address lost or delayed input events. Additionally, a drag event can be reissued by the remote device in frames during which no new input events are generated.