Image Encoding Motion Vector Modification for Flicker Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Flicker occurs in flat areas of images encoded by image encoding apparatuses due to decorrelation of prediction errors between frames, particularly when INTRA prediction modes or motion vectors differ between collocated blocks, leading to incorrect motion vectors and increased flicker visibility with larger block sizes.
Innovation Solution
The solution involves modifying motion vectors and prediction modes for blocks in flat areas by comparing encoding costs of different motion vectors and prediction modes, using threshold values based on quantization parameters and activity levels to select appropriate vectors and modes, thereby reducing flicker occurrences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If smaller quantization parameters are applied to still and flat areas, then flicker is reduced, but compression ratio deteriorates
Solution Approach 1:
The patent applies different quantization parameter adjustment strategies to different regions: only flat areas with flicker issues have their quantization parameters modified, while other regions maintain their original encoding parameters. This localized approach reduces flicker in problematic areas without unnecessarily degrading compression ratio across the entire image.
Solution Approach 2:
The patent applies quantization parameter adjustment only partially - specifically to flat areas where flicker occurs and where the activity level threshold is not exceeded. This partial application avoids the excessive compression ratio degradation that would result from applying the same adjustment to all image regions.
2Adaptability or versatility
If different INTRA prediction modes are selected for blocks in the same spatial location but in different frames, then prediction flexibility is improved, but flicker occurs due to decorrelation
Solution Approach 1:
The patent performs preliminary analysis of prediction error decorrelation between frames before finalizing prediction mode selection. By detecting decorrelation issues in advance, the system can adjust quantization parameters preemptively to prevent flicker, while still allowing prediction mode flexibility where it does not cause decorrelation problems.
Solution Approach 2:
The patent uses feedback from inter-frame comparison to detect when different prediction modes cause decorrelation and subsequent flicker. This feedback mechanism allows the system to identify problematic prediction mode combinations and adjust encoding parameters accordingly, balancing prediction flexibility with flicker prevention.
3Reliability
If motion vectors are changed for blocks in flat areas, then INTER flicker is reduced, but device complexity increases due to additional processing
Solution Approach 1:
The patent applies motion vector modification only to blocks located in flat areas where flicker is detected, rather than processing all blocks uniformly. This localized processing reduces the overall computational burden and device complexity while still achieving flicker reduction in the specific regions where it is needed.
Solution Approach 2:
The patent performs motion vector modification partially - only for blocks in flat areas with activity levels below the threshold and where flicker is detected. This partial application avoids the excessive processing complexity that would result from modifying motion vectors for all blocks in the frame.
Data Source
AI summary
There is provided an image encoding apparatus that divides an image in a video into blocks made up of a plurality of pixels, and encodes the blocks, the image encoding apparatus comprising: A first motion vector computation unit configured to compute a first motion vector of a block to be encoded in an image to be encoded; A motion vector modification unit configured to modify the first motion vector computed in the first motion vector computation unit to generate a second motion vector when a first condition is satisfied; Wherein the first condition is at least one of the following conditions: The absolute value of the first motion vector is greater than a first threshold The absolute value of the difference between the first motion vector and the candidate motion vector for modifying the first motion vector is greater than a second threshold.