VIDEO QUALITY ESTIMATION WITH A MACHINE LEARNING MODEL AS AN OPERATING SYSTEM SERVICE OR CLOUD SERVICE

Smart overview of the Invention

Video quality estimation is provided as an operating system or cloud service to collect quality estimates unobtrusively and without user feedback. This approach allows for consistent and reliable video quality assessments across different video streams, applications, and devices. By using machine learning models, the estimation process provides objective and detailed quality scores, enabling streaming and conferencing services to adjust video encoding for improved user experience.

Operating System Service

An operating system service on a client computer can estimate video quality by processing video data from applications. This service calculates quality metrics for reconstructed video content and maps them to a video quality score using a machine learning model. The results are then sent back to the application, allowing for consistent, timely, and unobtrusive feedback on video quality.

Application Interaction

Applications on client computers interact with the operating system service to receive video quality estimates. They provide video data to the service and receive estimation results, which can be used to generate feedback for cloud or streaming services. This interaction helps adjust subsequent video encoding, enhancing both video quality and user experience.

Cloud Service Operations

A cloud service on a server system estimates video quality by analyzing video data using a machine learning model. It calculates various quality metrics and generates encoder control values based on the estimation results. These values are sent to streaming or conferencing services, which use them to optimize encoding settings, ensuring consistent and timely quality assessments without client-side feedback.

Encoder Adjustments

Streaming or conferencing services utilize encoder control values from the cloud service to modify their encoding processes. By adjusting settings such as frame rate and bit rate according to these values, these services can enhance subsequent video quality and improve overall user satisfaction. This method leverages cloud-based insights to refine video delivery dynamically.