Video Camera Controller Frame-Type Noise Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional video cameras face a trade-off between improving video quality through better motion vector estimates and reducing video quality due to noise reduction, which often results in blur, especially when using inter-frame prediction techniques like MPEG-4 and H.264.

Innovation Solution

The method involves differentiating the capture and processing settings for image frames based on whether they will be encoded as intra-frames or inter-frames, with additional noise reduction applied only to intra-frames to improve motion vector accuracy without introducing blur in subsequent inter-frames, by modifying exposure time, gain, sharpness, and noise filtering settings.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If noise reduction is applied to image frames before encoding, then motion vector estimation accuracy is improved, but video quality deteriorates due to introduced blur

Engineering Contradiction:
Improvemotion vector estimation accuracyVSAvoidvideo quality degradation from blur
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent segments the video stream into different frame types (intra-frames and inter-frames) and applies different noise reduction strategies to each. Intra-frames receive aggressive noise reduction to improve motion vector estimation for subsequent inter-frames, while inter-frames use minimal or no noise reduction to preserve video quality and avoid blur artifacts.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by making the noise reduction strength frame-dependent rather than uniform across all frames. The processing quality is adapted locally to each frame's type and position in the GOP structure, optimizing the balance between motion vector accuracy and visual quality on a frame-by-frame basis.

Inventive Principle:
Principle #3Local quality

2Object-affected harmful factors

If aggressive noise reduction is applied to all frames, then video quality is improved by reducing noise, but motion vector estimation accuracy deteriorates due to excessive blur

Engineering Contradiction:
Improvenoise level in videoVSAvoidmotion vector estimation accuracy
Core Design Contradiction:
Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The patent performs preliminary aggressive noise reduction on intra-frames before they are used as reference frames for motion vector estimation in inter-frames. This preliminary action ensures that the reference frames have minimal noise, which improves the accuracy of motion vector estimation in subsequent inter-frames without requiring aggressive noise reduction on those frames.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses feedback from the encoder about which frames will be encoded as intra-frames or inter-frames to adjust the noise reduction processing. The encoder's frame type decisions feed back into the preprocessing stage, allowing the system to optimize noise reduction parameters based on the actual encoding requirements and GOP structure.

Inventive Principle:
Principle #23Feedback

3Loss of energy

If intra-frames are used as reference frames for inter-frame encoding, then bitrate is reduced through efficient compression, but video quality deteriorates if intra-frames contain noise

Engineering Contradiction:
Improvebitrate of encoded videoVSAvoidreference frame quality for motion estimation
Core Design Contradiction:
Loss of energyVSMeasurement precision

Solution Approach 1:

The patent segments the reference frame generation process by identifying intra-frames as special reference frames that require enhanced preprocessing. These segmented intra-frames receive dedicated aggressive noise reduction processing to ensure high quality reference images, while other frames use standard processing, thus maintaining low bitrate through efficient compression without sacrificing reference frame quality.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3379820B1Controller, video camera, and method for controlling a video camera
Publication Date: 2020.01.15 AXIS
  • EP3379820B1 patent drawingFigure 1
  • EP3379820B1 patent drawingFigure 2

AI summary

There is provided a method for controlling a video camera which is configured to capture and process image frames prior to encoding. The method comprises determining whether the video camera is going to encode an image frame as an intra-frame or an inter-frame. If the image frame is going to be encoded as an inter-frame, the video camera is controlled to capture and process the image frame using first settings prior to encoding. If the image frame is going to be encoded as an intra-frame, the video camera is controlled to capture and process the image frame using second settings prior to encoding, wherein the second settings are modified in relation to the first settings to further reduce a level of noise in the image frame.