Sampling-Based Super Resolution Video Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression methods, such as those using super resolution, face inefficiencies due to reliance on sub-pixel motion and fixed jitter strategies, which are not effective for non-static backgrounds, leading to suboptimal compression ratios and quality in video encoding and decoding.
Innovation Solution
A sampling-based super resolution approach that employs a smart down-sampling strategy capable of handling motion between frames, generating low-resolution frames and metadata to guide post-processing, allowing for improved super resolution reconstruction and increased compression ratios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sub-pixel motion is assumed for super resolution, then compression ratio is improved, but reliability deteriorates when sub-pixel motion is not present
Solution Approach 1:
The patent applies preliminary action by performing smart down-sampling on high-resolution frames before encoding to create low-resolution frames with enforced sub-pixel motion patterns. This preprocessing step ensures that sub-pixel motion information is embedded in advance, making it available for super resolution reconstruction without relying on the presence of natural sub-pixel motion in the original video sequence.
Solution Approach 2:
The patent changes parameters by transforming high-resolution frames into low-resolution frames with modified sampling patterns that enforce sub-pixel motion. The down-sampling process alters the spatial sampling parameters to create controlled sub-pixel shifts, and metadata is generated to track these parameter changes for subsequent reconstruction.
2Device complexity
If fixed jitter strategy is used for down-sampling, then device complexity is reduced, but adaptability deteriorates for non-static backgrounds
Solution Approach 1:
The patent applies dynamics by implementing a smart down-sampling strategy that adapts to motion between frames. Instead of using a fixed jitter pattern, the system dynamically adjusts the down-sampling process based on detected motion, allowing different regions of the frame to be sampled with appropriate shifts that account for both static and non-static content.
Solution Approach 2:
The patent applies segmentation by dividing the frame into different regions and applying different down-sampling strategies to each region based on motion characteristics. This allows the system to handle static backgrounds and moving foreground objects differently, improving overall adaptability while maintaining manageable complexity through region-based processing.
3Manufacturing precision
If high resolution video is directly encoded, then manufacturing precision is maintained, but loss of substance increases due to larger data size
Solution Approach 1:
The patent applies copying by creating low-resolution copies of high-resolution frames through smart down-sampling. These low-resolution copies contain essential visual information at reduced data size, and the original high-resolution information is preserved through metadata that guides reconstruction, effectively separating the compression function from the quality preservation function.
Solution Approach 2:
The patent applies the nested doll principle by embedding multiple levels of information: low-resolution frames contain the base visual content, while nested metadata contains motion information and sampling parameters. During reconstruction, these nested layers are combined to recover high-resolution quality, allowing progressive recovery of detail from compressed data.
Data Source
AI summary
Methods and apparatus are provided for sampling-based super resolution video encoding and decoding. The encoding method receives high resolution pictures and generates low resolution pictures and metadata there from, the metadata for guiding post-decoding post-processing of the low resolution pictures and the metadata; and then encodes the low resolution pictures and the metadata using at least one encoder. The corresponding decoding method receives a bitstream and decodes low resolution pictures and metadata there from using a decoder; and then reconstructs high resolution pictures respectively corresponding to the low resolution pictures using the low resolution pictures and the metadata.