Dynamic Exposure Adjustment for HDR Video Motion Artifacts

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing HDR image capture methods face challenges in balancing high dynamic range and sharp motion capture, particularly in video recording, as they often result in motion artifacts and require high frame rates, leading to inefficient exposure settings and increased noise in dark areas.

Innovation Solution

A method that dynamically adjusts exposure settings for successive frames based on both brightness distribution and motion analysis, using local and global motion data to minimize motion artifacts and optimize exposure without reducing frame rate, by setting exposures for each frame immediately after capturing the previous one, and merging frames to create HDR images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple frames are captured with varying exposures to cover full illumination distribution, then HDR image quality is improved, but motion artifacts increase and frame rate decreases

Engineering Contradiction:
ImproveHDR image qualityVSAvoidmotion artifacts
Core Design Contradiction:
Measurement precisionVSObject-generated harmful factors

Solution Approach 1:

The patent implements dynamic exposure adjustment where the exposure settings for successive frames are adapted based on detected motion between frames. The system calculates motion metrics from frame differences and modifies exposure parameters in real-time, transitioning from static predetermined exposures to dynamic motion-adaptive exposures, thereby reducing motion artifacts while maintaining HDR quality

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes exposure parameters (exposure time, gain) based on detected motion levels and brightness distribution characteristics. By adjusting these parameters dynamically according to scene conditions and motion detection results, the system optimizes the balance between capturing sufficient dynamic range and minimizing motion-related degradation

Inventive Principle:
Principle #35Parameter changes

2Illumination intensity

If exposure brackets are spaced further apart to cover full dynamic range, then HDR dynamic range is improved, but noise in dark areas increases

Engineering Contradiction:
Improvedynamic range coverageVSAvoidnoise level in dark areas
Core Design Contradiction:
Illumination intensityVSManufacturing precision

Solution Approach 1:

The patent applies different exposure weighting strategies to different spatial regions based on local brightness characteristics. Pixels in dark areas receive different weight factors compared to mid-tone and highlight regions during HDR merging, allowing optimized noise performance in dark areas while maintaining dynamic range coverage across the full image

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system uses feedback from analyzing the brightness distribution histogram and noise characteristics of captured frames to adjust exposure settings for subsequent frames. This closed-loop approach allows the system to optimize exposure bracket spacing and weighting based on actual scene conditions, reducing noise in dark areas while maintaining adequate dynamic range

Inventive Principle:
Principle #23Feedback

3Speed

If frame rate is increased to reduce motion blur, then motion sharpness is improved, but the number of usable frames for HDR decreases

Engineering Contradiction:
Improveframe rateVSAvoidnumber of usable HDR frames
Core Design Contradiction:
SpeedVSQuantity of substance

Solution Approach 1:

The system performs preliminary motion detection and brightness distribution analysis on incoming frames before committing them to HDR processing. By pre-evaluating motion metrics and exposure quality, the system can identify and select only those frames that are suitable for HDR merging, maximizing the utilization of captured frames at higher frame rates while maintaining HDR quality

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3046320B1Method for generating an HDR image of a scene based on a tradeoff between brightness distribution and motion
Publication Date: 2017.12.13 THOMSON LICENSING SA
  • EP3046320B1 patent drawingFigure 1
  • EP3046320B1 patent drawingFigure 2
  • EP3046320B1 patent drawingFigure 3

AI summary

When a current frame (fi) imaging said scene have been captured under a current exposure (Ei)), the method comprises: - setting an exposure (Ei+1) for a next frame (fi+1) as a function of the current exposure (Ei), of a first brightness distribution (Hfulli) of colors of pixels of the current frame (fi) and of a second brightness distribution (Hmotioni) of colors of those pixels weighted by the motion of those pixels, wherein said motion is evaluated for those pixels in comparison with a previous frame (fi), - capturing said next frame (fi+1) under said set exposure (Ei+1), - merging said next frame (fi+1) with at least said current frame (fi) into an HDR image.