Video Stabilisation Using Receiver-Side Motion Vectors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video stabilization methods require significant processing resources, making real-time stabilization challenging, especially in mobile devices where resources are limited and transmission is simultaneous with capture.

Innovation Solution

Stabilization is performed at the receive side of a communication system by decoding an encoded bit stream, determining camera motion, and compensating for motion using pixel displacement filtering and shifting, leveraging existing motion vectors in the bit stream to reduce computational complexity.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If image stabilisation is performed using motion estimation, motion smoothing and motion compensation blocks, then video stabilisation is achieved, but processing resources are significantly consumed

Engineering Contradiction:
Improvevideo stabilisationVSAvoidprocessing resources
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent inverts the traditional stabilization approach by performing stabilization at the receiver side rather than the camera side. The receiver uses motion vectors already present in the encoded bitstream to determine camera motion and applies compensation to the decoded video signal, avoiding the need for complex real-time processing at the camera while achieving the same stabilization effect

Inventive Principle:
Principle #13The other way round (Inversion)

Solution Approach 2:

The patent leverages motion vectors that are already calculated and embedded in the encoded video bitstream during compression. These pre-computed motion vectors are reused for stabilization purposes, eliminating the need to perform separate motion estimation and reducing processing requirements

Inventive Principle:
Principle #10Preliminary action

2Productivity

If motion vectors are used to encode the video signal, then encoding efficiency is improved, but computational complexity increases at the receive side for stabilization

Engineering Contradiction:
Improveencoding efficiencyVSAvoidcomputational complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent makes the motion vectors embedded in the bitstream serve dual purposes: both for video decoding reconstruction and for camera motion estimation for stabilization. The same motion information that enables efficient decoding also provides the basis for stabilization, eliminating redundant computation

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The motion vectors in the encoded bitstream are utilized for multiple functions: they serve both as decoding parameters for reconstructing video frames and as motion indicators for determining camera movement and applying stabilization compensation

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP2759124B1Received video stabilisation
Publication Date: 2017.01.11 SKYPE
  • EP2759124B1 patent drawing
  • EP2759124B1 patent drawing
  • EP2759124B1 patent drawing

AI summary

Method, device and computer program product for stabilising a received video signal. A plurality of frames of the video signal are captured using a camera. The video signal is transmitted as an encoded bit stream. Displacement of the camera between successive frames is determined from a motion indication in the bit stream. The determined displacement is used to compensate for motion in the video signal between successive frames caused by the motion of the camera, to thereby stabilise the video signal.