Motion Compensated Video Spatial Up-Conversion for Artifact Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video spatial up-conversion techniques fail to effectively address aliasing and over-smoothness artifacts, particularly in the vertical direction, due to the lack of proper anti-aliasing filtering and neglect of temporal correlations in motion trajectories, resulting in blurring of video signals.
Innovation Solution
The implementation of motion-compensated video up-conversion techniques that differentiate between horizontal and vertical interpolation, utilizing motion-compensated methods for vertical interpolation, and leveraging the connection between video spatial up-conversion and deinterlacing to enhance spatial resolution while suppressing aliasing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional spatial interpolation is used for video spatial up-conversion, then the spatial resolution is enhanced, but aliasing and over-smoothness artifacts occur due to lack of proper anti-aliasing filtering
Solution Approach 1:
The patent segments the up-conversion process into separate horizontal and vertical interpolation stages. Horizontal interpolation is performed first on the original low-resolution video, then vertical interpolation is performed on the horizontally up-sampled video. This segmentation allows different filtering strategies to be applied to each direction, with motion-compensated filtering specifically applied to vertical interpolation to suppress aliasing artifacts while preserving high-frequency components.
Solution Approach 2:
The patent applies motion compensation as a preliminary action before vertical interpolation. Motion vectors are estimated from reference frames, and pixel values are predicted along motion trajectories before the actual interpolation occurs. This preliminary motion-compensated prediction prepares the data in a way that reduces aliasing and preserves high-frequency information during the subsequent vertical up-sampling process.
2Manufacturing precision
If conventional spatial interpolation is used for video spatial up-conversion, then the spatial resolution is enhanced, but over-smoothness artifacts occur due to suppression of high-frequency components
Solution Approach 1:
Motion compensation is performed as a preliminary action that predicts pixel values along motion trajectories before interpolation. This prediction preserves high-frequency components by utilizing temporal correlations from reference frames, preventing the loss of fine details that would otherwise occur during up-sampling. The motion-compensated predictions are then used to guide the interpolation process, ensuring high-frequency information is retained.
Solution Approach 2:
The patent changes the filtering parameters adaptively based on motion information. Motion vectors and motion compensation results are used to adjust the interpolation filters, allowing the system to preserve high-frequency components in regions with motion while maintaining smooth interpolation in stationary regions. This parameter adaptation prevents over-smoothness by preserving edge and detail information.
3Object-generated harmful factors
If motion-compensated methods are applied to vertical interpolation, then aliasing is suppressed and high-frequency components are preserved, but computational complexity increases
Solution Approach 1:
The patent segments the computational workload by applying motion compensation only to vertical interpolation, while horizontal interpolation uses simpler filtering methods. This selective application reduces overall computational complexity compared to applying motion compensation to both directions. The segmentation allows the system to focus computational resources on the direction where aliasing is most problematic while maintaining efficiency in the horizontal direction.
Solution Approach 2:
The patent applies different quality levels of processing to different directions. Vertical interpolation receives full motion-compensated treatment with adaptive filtering to suppress aliasing and preserve high-frequency components, while horizontal interpolation uses standard filtering methods. This local quality differentiation optimizes the balance between processing quality and computational complexity by applying sophisticated methods only where most needed.
Data Source
AI summary
A method for performing motion compensated video spatial up-conversion on video. The horizontal samples in successive fields are first interpolated using a spatial interpolation technique. This is followed by interpolating the corresponding vertical samples using a motion compensated deinterlacing technique. Such techniques can include an adaptively recursive motion compensated video spatial up-conversion or an adaptively recursive motion compensated video spatial up-conversion using a generalized sampling theorem. The present invention can be used to convert video captured on a mobile device, such as a mobile telephone, so that it can be subsequently and adequately displayed on a television.