Entropy Coding for Video Data Efficiency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Conventional entropy coding techniques are inefficient due to the requirement to transmit codewords indicating the absence of zero-valued coefficients between consecutive nonzero coefficients, leading to inefficiencies, especially when long sequences of nonzero values occur.

Innovation Solution

The proposed entropy coding method transmits a count of consecutive nonzero values followed by their values, then a count of consecutive zero values, avoiding the transmission of codewords indicating the absence of zero-valued coefficients between them.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional entropy coding techniques transmit codewords indicating the absence of zero-valued coefficients between consecutive nonzero coefficients, then the coding process maintains simplicity and uniformity, but coding efficiency deteriorates due to unnecessary transmissions

Engineering Contradiction:

Improvecoding efficiencyVSAvoidtransmission data volume

Core Design Contradiction:

ProductivityVSLoss of substance

Solution Approach 1:

The patent extracts and eliminates the unnecessary codewords that indicate the absence of zero-valued coefficients between consecutive nonzero coefficients. By removing these redundant transmissions from the coding process, the patent achieves higher coding efficiency without sacrificing the ability to accurately represent the video data.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent discards the conventional approach of transmitting codewords for absent zero-coefficients and recovers efficiency by only transmitting necessary information. The decoder recovers the original data structure by understanding that omitted codewords imply the presence of consecutive nonzero coefficients, eliminating waste while preserving information.

Inventive Principle:

Principle #34Discarding and recovering

2Ease of operation

If conventional entropy coding transmits data identifying the number of zero-valued coefficients followed by the first nonzero coefficient, then the coding structure remains consistent and simple to decode, but transmission overhead increases when long sequences of nonzero values occur

Engineering Contradiction:

Improvedecoding simplicityVSAvoidtransmission time

Core Design Contradiction:

Ease of operationVSLoss of time

Solution Approach 1:

The patent inverts the conventional coding approach by transmitting data identifying the number of nonzero coefficients first, followed by the nonzero coefficient values themselves, rather than coding zero-coefficients first. This inversion eliminates unnecessary transmissions when long sequences of nonzero values occur, reducing transmission time while maintaining decoding simplicity through the consistent use of count-value pairs.

Inventive Principle:

Principle #13The other way round (Inversion)

Data Source

PatentUS10873765B2Techniques for high efficiency entropy coding of video data

Publication Date: 2020.12.22 APPLE INC

AI summary

Entropy coding/decoding techniques are disclosed in which data is coded alternately as a series of nonzero values and zero values until the transmitted data is consumed. Nonzero values may be coded first with transmission of data identifying a number of consecutive nonzero values that appear in scan order followed by transmission of the nonzero values themselves. Thereafter, if other data remains to be transmitted, data may be transmitted identifying a number of consecutive zero values that appear next in scan order followed by transmission of a next nonzero value encountered in scan order. By transmitting the nonzero values as a group, it is expected that the proposed entropy-coding process will achieve higher efficiency than competitive techniques.