18477493. REALISTIC DISTRACTION AND PSEUDO-LABELING REGULARIZATION FOR OPTICAL FLOW ESTIMATION simplified abstract (QUALCOMM Incorporated)

From WikiPatents
Jump to navigation Jump to search

REALISTIC DISTRACTION AND PSEUDO-LABELING REGULARIZATION FOR OPTICAL FLOW ESTIMATION

Organization Name

QUALCOMM Incorporated

Inventor(s)

Jisoo Jeong of San Diego CA (US)

Risheek Garrepalli of San Diego CA (US)

Hong Cai of San Diego CA (US)

Fatih Murat Porikli of San Diego CA (US)

REALISTIC DISTRACTION AND PSEUDO-LABELING REGULARIZATION FOR OPTICAL FLOW ESTIMATION - A simplified explanation of the abstract

This abstract first appeared for US patent application 18477493 titled 'REALISTIC DISTRACTION AND PSEUDO-LABELING REGULARIZATION FOR OPTICAL FLOW ESTIMATION

Simplified Explanation

The computer-implemented method described in the abstract involves generating augmented frames by combining images and frames, estimating optical flow, and updating parameters based on loss between the estimation and a target.

  • Combining images and frames to generate augmented frames
  • Estimating optical flow using a model
  • Updating parameters or weights of the model based on loss between estimation and target

Potential Applications

This technology could be used in:

  • Video editing software
  • Virtual reality applications
  • Surveillance systems

Problems Solved

This technology helps to:

  • Improve video quality
  • Enhance motion tracking
  • Optimize image processing

Benefits

The benefits of this technology include:

  • Enhanced visual effects
  • Improved accuracy in motion estimation
  • Efficient video processing

Potential Commercial Applications

The potential commercial applications of this technology include:

  • Entertainment industry for special effects
  • Security industry for surveillance systems
  • Technology companies for image processing software

Possible Prior Art

One possible prior art for this technology could be:

  • Existing optical flow estimation models
  • Image processing algorithms

What are the limitations of this technology in real-time applications?

In real-time applications, the computational complexity of the optical flow estimation model may lead to delays in processing speed, affecting the real-time performance of the system.

How does this technology compare to traditional methods of optical flow estimation?

This technology improves upon traditional methods by incorporating augmented frames to enhance the accuracy of flow estimation, leading to better results in various applications such as video editing and motion tracking.


Original Abstract Submitted

A computer-implemented method includes generating a first augmented frame by combining a first image and a first frame of a first frame pair. The computer-implemented method also includes generating, via an optical flow estimation model, a first flow estimation based on a second frame of the first frame pair and the first augmented frame. The computer-implemented method further includes updating one or both of parameters or weights of the optical flow estimation model based on a first loss between the first flow estimation and a training target.