Invention Title:

ADAPTIVE FOVEATED ENCODER AND GLOBAL MOTION PREDICTOR

Publication number:

US20250280147

Publication date:

2025-09-04

Section:

Electricity

Class:

H04N19/527

Inventors:

Ce Wang 🇺🇸 Bellevue, WA, United States

CHANGLIANG WANG 🇺🇸 Bellevue, WA, United States

Yongfa Zhou 🇨🇳 Beijing, China

Yu Yang 🇨🇳 Beijing, China

Ping LIU 🇺🇸 Sunnyvale, CA, United States

Yunbiao Lin 🇨🇳 Shanghai, China

Bo Qiu 🇺🇸 Hillsboro, OR, United States

ZHAN LOU 🇨🇳 Shanghai, China

Jianwei Yang 🇨🇳 Beijing, China

Yating Wang 🇨🇳 Beijing, China

Bo Zhao 🇨🇳 Shanghai 31, China

Wenyi Tang 🇨🇳 Beijing 11, China

Assignee:

INTEL CORPORATION 🇺🇸 Santa Clara, CA, United States

Applicant:

Intel Corporation 🇺🇸 Santa Clara, CA, United States

Smart overview of the Invention

The patent application discusses an adaptive video encoder that adjusts video encoding parameters based on information from a headset. This includes focus-related and motion-related data, which are used to optimize the encoding of video images. The technology is particularly relevant to graphics systems and introduces an adaptive foveated encoder and global motion predictor, enhancing image processing by focusing on areas where visual acuity is highest.

System Components

The described system includes several key components: an application processor, persistent storage media, a graphics subsystem, and various engines for sensing, focus, and motion. These components work together to provide sensed information, focus data, and motion data. The adaptive encoder uses this information to adjust video encoding parameters dynamically. The system's architecture allows for flexible implementation in hardware or software, utilizing technologies like FPGAs or ASICs.

Sense Engine

The sense engine gathers data from multiple sources such as cameras, gyroscopes, accelerometers, and biometric sensors. This data provides comprehensive sensed information that can include image, audio, and motion details. The sense engine plays a crucial role in adjusting video encode parameters by integrating input from various user devices like head-mounted displays (HMDs) or smartphones.

Focus Engine

The focus engine processes data from the sense and motion engines to determine the user's focal point and related parameters like eye position and pupil size. This information helps in adjusting video encoding to align with where the user is looking or expected to look. By leveraging eye tracking technology, the system can enhance user experience by focusing on regions of interest within the visual field.

User Interaction and Application

In practical terms, when a user wears a VR headset equipped with this technology, the system can dynamically adjust the focus based on real-time eye tracking data. It considers factors like pupil dilation and depth of focus to determine the user's focal area accurately. This capability is crucial for applications requiring high visual fidelity and user immersion, such as virtual reality environments.