Low Resolution Motion Estimation Pipeline for Video Encoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video encoding systems face challenges in efficiently encoding image data, particularly for higher resolution images, due to the computational complexity of determining encoding parameters, which can hinder real-time or near real-time transmission and display.

Innovation Solution

The implementation of a low resolution pipeline in parallel with the main encoding pipeline, which determines encoding parameters using low-resolution image data derived from source image data, allowing for reduced memory bandwidth and improved operational efficiency by performing motion estimation searches ahead of the main pipeline.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If full-resolution image data is used for motion estimation searches, then motion estimation accuracy is improved, but computational complexity and processing time increase significantly

Engineering Contradiction:
Improvemotion estimation accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides the motion estimation process into two segments: a first motion estimation search performed on downsampled/low-resolution image data to identify candidate motion vectors, and a second motion estimation search performed on full-resolution image data using only those candidate vectors. This segmentation reduces the overall computational complexity while maintaining motion estimation accuracy.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs a preliminary motion estimation search on downsampled image data before the main motion estimation search on full-resolution data. This preliminary action identifies candidate motion vectors that narrow down the search space for the subsequent full-resolution search, reducing computational complexity while preserving accuracy.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If full-resolution image data is used for motion estimation searches, then motion estimation accuracy is improved, but processing time increases

Engineering Contradiction:
Improvemotion estimation accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the motion estimation process into two time-efficient stages: a fast preliminary search on downsampled data that quickly identifies candidate vectors, and a focused second search on full-resolution data using only those candidates. This segmentation significantly reduces total processing time compared to performing a single exhaustive search on full-resolution data.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs a preliminary motion estimation search on downsampled image data before the main motion estimation search. This preliminary action reduces the search space for the subsequent full-resolution search, thereby reducing processing time while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

3Device complexity

If downsampled image data is used for motion estimation searches, then computational complexity is reduced, but motion estimation accuracy deteriorates

Engineering Contradiction:
Improvecomputational complexityVSAvoidmotion estimation accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent uses downsampled image data for the first motion estimation search to reduce computational complexity, then performs a second motion estimation search on full-resolution image data using the candidate motion vectors identified in the first search. This two-stage segmentation ensures that accuracy is restored in the final step while benefiting from reduced complexity in the preliminary step.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs a preliminary motion estimation search on downsampled image data to identify candidate motion vectors, which are then used to guide the second search on full-resolution data. This preliminary action on reduced-complexity data does not compromise final accuracy because the candidate vectors are refined in the second search stage.

Inventive Principle:
Principle #10Preliminary action

4Measurement precision

If multiple motion estimation searches are performed, then motion estimation accuracy is improved, but processing time increases

Engineering Contradiction:
Improvemotion estimation accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent divides the multiple motion estimation searches into two distinct segments: a first search on downsampled data with a larger search range to identify candidates, and a second search on full-resolution data with a focused search range using the candidates. This segmentation optimizes the balance between accuracy and processing time by avoiding redundant searches.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs a preliminary motion estimation search that identifies candidate motion vectors, which are then used to constrain and guide the second search. This preliminary action reduces the effective search space for the second search, thereby reducing processing time despite performing multiple searches.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS12212776B2Systems and methods for low resolution motion estimation searches
Publication Date: 2025.01.28 APPLE INC
  • US12212776B2 patent drawing
  • US12212776B2 patent drawing
  • US12212776B2 patent drawing

AI summary

A video encoding system encodes source image data corresponding with an image includes a low resolution pipeline that receives the source image data corresponding with a first coding block in the image. The low resolution pipeline includes a low resolution motion estimation block programmed to generate a first downscaled coding block by downscaling resolution of the source image data corresponding with the first coding block. The first downscaled coding block comprises a first downscaled prediction block corresponding with a first prediction block in the first coding block. The low resolution pipeline may also perform several low resolution motion estimation searches to generate motion vector candidates. The video encoding system also includes a main pipeline that receives the source image data and determines encoding parameters to be used to encode the first coding block based at least partially on the motion vector candidates.