The purpose is to remove spatial redundancy and temporal redundancy. Because there are a lot of spatial redundancy and temporal redundancy in the video, including spatial redundancy and temporal redundancy, after prediction, the residual error that removes most of the spatial or temporal redundancy is obtained.
Neighboring pixels in the image space have a strong correlation, and intra-frame prediction technology removes spatial redundancy.
Video images have a strong correlation in time, that is, there is temporal redundancy. Coding techniques for removing temporal redundancy include: Motion Estimation (ME) and Motion Compensation (MC).
The purpose is to remove temporal redundancy of video images. Motion estimation searches for the best matching prediction block for the current block within the search range, and the motion estimation of the full search method has high computational complexity.
To reduce the number of searches for motion estimation while maintaining prediction accuracy, typical algorithms are:
Three Step Search (Three Step Search, TSS)
Two-dimensional Log Search (2D Logarithmic Search, 2DLOG)
Orthogonal Search Algorithm (OSA)
Cross Search Algorithm (CSA)
New Three Step Search (New Three Step Search, NTSS)
Four Step Search (Four Step Search, FSS)
Conjugate Direction Search (CDS)
Gradient Descent Search (GDS)
Hierarchical Block Matching Algorithm (HBMA)
The temporal motion position is more likely to be between integer pixels, that is, sub-pixels. The value of the sub-pixel can be estimated by using adjacent integer pixels: the value of the sub-pixel is usually obtained by linear or bilinear interpolation. Sub-pixel motion estimation has higher prediction accuracy, but the complexity is also higher: 1/2 sub-pixel motion estimation, image storage space increased by 4 times, motion vector needs to be enlarged by 2 times, 1/4 sub-pixel motion estimation , the image storage space is increased by 16 times, the motion vector needs to be enlarged by 4 times, and the computational complexity is also doubled. Instant messaging chat software app development can add Wei Keyun's v: weikeyun24 consultation
The motion of video content is very complex, and image block coding can better improve motion prediction accuracy and compression efficiency. There is a trade-off between the coded block size and the number of coded bits of the additional information (MV, Mode). Smaller coded block sizes have better predictions but more bits of additional information.
Newly appearing objects refer to future frames to have a better prediction effect, and the average of the two predictions before and after can reduce the prediction variance.
Based on a global affine motion model. The prediction accuracy is not as good as block-based motion estimation. The number of MVs is small, and it is suitable for motion estimation of simple motion scenes.