Data Frame Encoding Using Predictive Transform Subtraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital signal processing techniques for encoding and decoding sequences of data frames, such as video frames, require significant computational resources, memory, and power due to the complexity of transformation and prediction methods, particularly in processes like discrete cosine transformation (DCT) and quantization.
Innovation Solution
The approach involves applying a transformation with distributive properties, such as integer DCT, to both original and predicted data frames, minimizing the generation of different values and reducing resource requirements by using mean or vertical prediction modes that simplify the transformation and subtraction processes, thereby reducing computational and memory needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If complex transformation and prediction methods (such as DCT) are used for encoding data frames, then encoding quality and compression efficiency are improved, but computational resources, memory, and power consumption increase significantly
Solution Approach 1:
The patent extracts and processes only the essential different values between original and predicted data frames, rather than processing all data elements. By identifying and isolating only the non-zero difference values that contain actual information, the method eliminates redundant processing of identical or predictable values, thereby reducing computational load and power consumption while maintaining encoding quality
Solution Approach 2:
The patent applies partial action by performing transformation operations only on the necessary portions of data - specifically, only on the difference values that are non-zero. Instead of applying DCT or other complex transformations to entire data frames, the method selectively processes only the minimal required data elements, reducing overall computational resources and energy usage
2Measurement precision
If complex transformation and prediction methods are used for encoding data frames, then encoding quality is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent extracts only the essential difference values between original and predicted frames for processing. By separating and processing only the non-zero difference elements, the method reduces computational complexity while maintaining the accuracy needed for faithful reconstruction of the original data during decoding
Solution Approach 2:
The patent performs prediction operations in advance to generate predicted data frames before the actual encoding process. This preliminary action identifies which values will be different, allowing the subsequent transformation and encoding steps to focus only on those specific difference values, thereby reducing overall computational complexity while preserving decoding accuracy
3Loss of information
If traditional encoding methods processing all data elements are used, then complete data representation is achieved, but memory requirements and processing overhead increase
Solution Approach 1:
The patent extracts and stores only the non-zero difference values between original and predicted frames, rather than storing or processing all data elements. This selective extraction maintains data completeness for the actual information content while significantly reducing memory requirements by eliminating storage of redundant identical values
Solution Approach 2:
The patent applies partial processing by operating only on the necessary subset of data - the non-zero difference values. This approach maintains complete representation of actual information while using minimal memory resources, as only the essential varying elements are processed and stored rather than all possible data elements
Data Source
AI summary
Encoding data by first performing a transformation of predicted data and input data, and then performing a subtraction of the resulting outputs. In an embodiment, the prediction approach is chosen such that fewer elements of different values (compared to a number of elements in the input data) are generated, and the different values are generated in a predictable position. The transformation approach is chosen such that the output expressly represents variations in the input data as well as satisfies a distributive property. The decoding may be performed based on the same concepts. As a result, the data can be encoded and/or decoded efficiently.