Motion Vector Encoding Omission for Bit Overhead Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The H.264/AVC video compression standard generates a large number of motion vectors due to variable block size segmentation, leading to increased bit overhead and reduced effectiveness of motion compensation, despite decreased residual data.
Innovation Solution
A method for encoding motion vector information that calculates Predictive Motion Vectors (PMVs) for subunits, determines if they satisfy a predetermined condition, and generates coded motion vector data that omits unnecessary motion vectors if the condition is met, using a controller and motion estimator to manage the encoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If variable block size segmentation is used to improve motion compensation accuracy, then residual data decreases, but the number of motion vectors increases leading to increased bit overhead
Solution Approach 1:
The patent extracts only the necessary motion vector information by comparing PMVs with actual motion vectors and identifying when they are equal. When equality is detected, the entire motion vector can be omitted from encoding, extracting only the essential information needed for accurate reconstruction while discarding redundant data.
Solution Approach 2:
The patent changes the encoding parameter from always transmitting full motion vectors to conditionally transmitting based on equality with PMVs. This parameter change allows the encoder to switch between transmitting full motion vectors and omitting them entirely, optimizing the bit rate based on the actual difference between predicted and actual motion vectors.
2Manufacturing precision
If more subunits are created to improve motion compensation precision, then residual data decreases, but the number of motion vectors increases
Solution Approach 1:
The patent extracts only the essential motion vector information by detecting when PMVs equal actual motion vectors. When equality is detected, entire motion vectors are removed from the encoded data stream, reducing the quantity of motion vectors that need to be transmitted while preserving reconstruction accuracy.
Solution Approach 2:
The patent discards motion vector information when it can be accurately recovered from PMVs alone. By detecting equality between PMVs and actual motion vectors, the encoder can discard the actual motion vector data, knowing it can be perfectly recovered at the decoder using the same PMV calculation, thus reducing the number of transmitted vectors.
3Reliability
If all motion vectors are encoded to ensure accurate recovery, then video quality is maintained, but bit rate increases
Solution Approach 1:
The patent uses feedback by comparing PMVs with actual motion vectors and using this comparison result to control encoding. The equality detection creates a feedback mechanism where the encoder adjusts its output based on whether PMVs sufficiently represent the actual motion vectors, maintaining quality only when necessary and reducing bit rate when PMVs are adequate.
Solution Approach 2:
The patent changes the encoding parameter from always encoding full motion vectors to conditionally encoding based on equality with PMVs. This parameter change allows the system to maintain video quality when PMVs are insufficient while reducing bit rate when PMVs are adequate, optimizing the trade-off between quality and compression.
Data Source
AI summary
A method and apparatus for encoding/decoding motion vector information are provided, in which a coding unit is divided into a plurality of subunits, PMVs of the subunits are calculated, it is determined whether the PMVs satisfy a predetermined condition, and coded motion vector data are generated, which include information indicating an omission of the motion vectors of the subunits, if the PMVs satisfy the predetermined condition.