Intra mode selection method for skipping large-size CU
A technology of mode selection and size, applied in the field of video processing, can solve the problems of long encoding time and large amount of calculation, and achieve the effect of saving encoding time and speeding up encoding.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0050] In the prior art, methods for selecting the Intra mode include:
[0051] Step 1, the Intra mode selection of a CU in the I frame starts;
[0052] Step 2, do the prediction of the current Intra mode,
[0053] Step 3, calculate the Sad of the difference between the predicted value and the pixel value of this CU, and put it into a list of Sad; all Intra modes are completed, and then execute step 4, otherwise return to step 2;
[0054] Step 4. Arrange all the values in the Sad list in ascending order. Take the first N Intra modes and use a more accurate and more computationally intensive method to select an optimal Intra mode as the mode selection result of this CU. This CU has a multi-sub-CU division mode specified in the standard, and the Intra mode of one CU in an I frame Select End.
Embodiment 2
[0056] The difference from Embodiment 1 is that this embodiment provides an Intra mode selection method that skips a large-size CU, which is applied to an I frame of a VVC encoder, and the method includes:
[0057] S1, judge the height or width of the CU, if the height or width is greater than 64, execute S3, otherwise execute S2;
[0058] S2, judge the height or width of the CU, when the height or width is greater than 32, and the Qp is less than the first Qp threshold; then execute S3, otherwise execute S6;
[0059]S3, calculate the variance value Var between CU pixels; and judge the variance value, when Var is smaller than the variance threshold VarThr; then execute S4, otherwise execute S11;
[0060] S4. Predict the CU in at least three directions, and determine the sad value of each direction according to the predicted value of each direction prediction and each pixel value, put the sad value of each direction into the Sad list, and determine the minimum predicted value ...
Embodiment 3
[0087] On the basis of Embodiment 2, the difference from Embodiment 2 is that this embodiment replaces vertical mode prediction with vertically adjacent direction prediction, such as Intra_ANGULAR49 or NTRA_ANGULAR51; horizontal mode prediction is replaced with horizontally adjacent Either direction forecast, such as NTRA_ANGULAR17 or NTRA_ANGULAR19.
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com