Reduced Complexity Entropy Coding Partition Function
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing entropy coding methods require complex hardware and high processing power due to large bit usage, leading to increased latency and power consumption, while also compromising compression accuracy with approximation errors.
Innovation Solution
The implementation of reduced complexity entropy coding, which partitions intervals into sub-intervals based on probability using a reduced complexity partition function with a power of two reduction value, accumulating approximation errors at the first sub-interval to simplify decoding and reduce hardware requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing entropy coding methods are used with large bit usage, then compression accuracy is maintained, but hardware complexity and processing power requirements increase
Solution Approach 1:
The patent changes the parameter of bit usage in the partition function from large values to power-of-two values (e.g., 2^k). This parameter change simplifies the hardware implementation by enabling the use of shift operations instead of complex division operations, thereby reducing hardware complexity while maintaining acceptable compression accuracy through error accumulation and compensation mechanisms.
2Measurement precision
If existing entropy coding methods are used with large bit usage, then compression accuracy is maintained, but processing speed decreases due to increased latency
Solution Approach 1:
The patent changes the parameter of the partition function to use power-of-two values, which enables faster computation through shift operations. This parameter change directly improves processing speed by reducing the computational complexity of the partitioning operation, while compression accuracy is maintained through the error accumulation mechanism that compensates for the reduced precision.
3Measurement precision
If existing entropy coding methods are used with large bit usage, then compression accuracy is maintained, but power consumption increases
Solution Approach 1:
The patent changes the partition function parameter to power-of-two values, which reduces power consumption by simplifying the computational operations required. The use of shift operations instead of division operations significantly reduces the energy required for each encoding operation, while the error accumulation mechanism ensures that compression accuracy is maintained despite the reduced computational precision.
4Speed
If reduced complexity partition function is used with power of two reduction value, then processing speed and hardware efficiency are improved, but compression accuracy may be compromised
Solution Approach 1:
The patent segments the compression process into two parts: the partitioning operation using the simplified power-of-two partition function for speed, and the error accumulation mechanism to preserve accuracy. By separating these functions, the system can use fast shift operations for partitioning while compensating for the resulting precision loss through systematic error tracking and accumulation, thus resolving the contradiction between speed and accuracy.
Solution Approach 2:
The patent implements a feedback mechanism through error accumulation, where the approximation errors introduced by the simplified partition function are tracked and compensated for in subsequent operations. This feedback loop ensures that the compression accuracy is maintained despite the use of the reduced-complexity partition function, allowing the system to achieve both speed and accuracy goals.
Data Source
AI summary
Content to be encoded may be defined by a sequence of values chosen from a set of possible values. Individual values of the set of possible values may have a probability of occurrence in the sequence of values. The set of possible values may have a total probability value. The content may be encoded based on a reduced complexity entropy coding and an interval. The reduced complexity entropy coding may partition the interval into sub-intervals based the probability of individual values that appears in the sequence of values and a reduced complexity partition function. The reduced complexity partition function may have an interval size value reduced by a reduction value less than the total probability value.