Nvidia corporation (20240114144). FRAME SELECTION FOR STREAMING APPLICATIONS simplified abstract
Contents
- 1 FRAME SELECTION FOR STREAMING APPLICATIONS
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 FRAME SELECTION FOR STREAMING APPLICATIONS - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Unanswered Questions
- 1.11 Original Abstract Submitted
FRAME SELECTION FOR STREAMING APPLICATIONS
Organization Name
Inventor(s)
Aurobinda Maharana of Chinchwad (IN)
Vignesh Ungrapalli of Udupi (IN)
Ming-Yu Liu of San Jose CA (US)
FRAME SELECTION FOR STREAMING APPLICATIONS - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240114144 titled 'FRAME SELECTION FOR STREAMING APPLICATIONS
Simplified Explanation
The patent application addresses reference frame selection in video streaming applications by using processing units to identify blurred frames based on variance of motion.
- Explanation of the patent:
* The system uses processing units to analyze frames in a video stream. * Frames with a variance of motion below a certain threshold are identified as blurred frames. * The threshold is adaptive and based on a moving average of variance of motion determined using reference frames.
Potential Applications
This technology could be applied in:
- Video streaming services
- Video surveillance systems
- Virtual reality applications
Problems Solved
This technology helps in:
- Improving video quality by identifying and filtering out blurred frames
- Enhancing user experience by reducing visual disturbances in video streams
Benefits
The benefits of this technology include:
- Enhanced video quality
- Improved accuracy in motion detection
- Optimized video streaming performance
Potential Commercial Applications
- Video streaming platforms for improved video quality
- Security companies for enhanced surveillance systems
- Virtual reality companies for better user experiences
Possible Prior Art
One possible prior art could be the use of motion detection algorithms in video processing systems to identify blurry frames.
Unanswered Questions
How does the system handle different types of motion blur in video frames?
The patent abstract does not specify how the system distinguishes between different types of motion blur, such as linear blur or rotational blur.
What is the computational overhead of implementing this system in real-time video streaming applications?
The abstract does not provide information on the computational resources required to implement this system in real-time video streaming applications.
Original Abstract Submitted
systems and methods herein address reference frame selection in video streaming applications using one or more processing units to identify a frame of a sequence of frames as a blurred frame based at least in part on a first variance of motion (vom) of the frame being less than or equal to an adaptive threshold that is based in part on a moving average of variance of motion (maov) determined using one or more reference frames.
