Real-Time Virtual Object Motion Blur Using Device Sensor Data

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods fail to effectively blur virtual objects in real-time during video acquisition by moving devices, resulting in a lack of realism in augmented reality applications, as existing techniques are computationally intensive and not directly applicable to capturing devices.

Innovation Solution

Estimating an apparent motion vector between successive images captured by a device, using motion data from motion sensors, to filter and blur virtual objects in real-time, improving the realism of augmented reality applications on mobile devices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If post-production blurring methods are used for virtual objects, then motion blur realism is improved, but computational resources and processing time are excessively consumed

Engineering Contradiction:
Improvemotion blur realismVSAvoidprocessing speed
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent applies preliminary action by estimating the motion blur parameters in advance based on device motion sensors (gyroscope, accelerometer) before the actual rendering of virtual objects. This allows the blurring effect to be pre-calculated using sensor data that reflects the camera's movement, enabling real-time application without heavy post-production computation. The motion vector and blur kernel are determined preliminarily from sensor readings, so when virtual objects are rendered, the blur can be applied efficiently using these pre-computed parameters.

Inventive Principle:
Principle #10Preliminary action

2Manufacturing precision

If existing blurring methods are applied to virtual objects captured by moving devices, then motion blur is added, but the integration with real scene motion becomes inconsistent

Engineering Contradiction:
Improvemotion blur consistencyVSAvoidmethod applicability
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent merges the blurring process for virtual objects with the motion characteristics of the real scene by using the same device motion sensors (gyroscope, accelerometer) to determine the motion parameters for both. The apparent motion vector is calculated by combining the device's rotational and translational motion data, ensuring that virtual objects receive the same motion blur treatment as the real scene. This unified approach using combined sensor data ensures consistency between virtual and real elements in the augmented reality composition.

Inventive Principle:
Principle #5Merging (Combining)

3Productivity

If real-time blurring is implemented without motion sensor data, then processing speed is maintained, but motion blur accuracy and realism are significantly reduced

Engineering Contradiction:
Improvereal-time processing capabilityVSAvoidmotion blur accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent introduces motion sensors (gyroscope, accelerometer) as intermediaries to bridge the gap between real-time processing requirements and accurate motion blur generation. These sensors provide continuous motion data that mediates between the device's physical movement and the computational blurring process. The sensor data serves as an intermediary input that translates physical camera motion into accurate motion vectors and blur kernels, enabling real-time processing with high precision without requiring complex post-production analysis.

Inventive Principle:
Principle #24Intermediary (Mediator)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

Enables real-time blurring of virtual objects in videos captured by moving devices, enhancing the realism and efficiency of augmented reality applications by leveraging device motion data for accurate motion blur simulation.

Implementation Method 1

said motion of said device is obtained from an angular rate generated by at least one motion sensor of said device

Methodology Applied
Scientific EffectAngular rate measurement: Gyroscope

Data Source

PatentEP3296952B1Method and device for blurring a virtual object in a video
Publication Date: 2020.11.04 INTERDIGITAL CE PATENT HOLDINGS SAS
  • EP3296952B1 patent drawingFigure 1
  • EP3296952B1 patent drawingFigure 2
  • EP3296952B1 patent drawingFigure 3~4

AI summary

In order to blur a virtual object in a video in real time as the video is acquired by a device capturing a real scene, a salient idea comprises estimating an apparent motion vector between two successive images, being captured at two successive device poses, wherein the apparent motion vector estimation is based on a motion of the device. The successive images are then filtered based on the estimated apparent motion vector.