Fast QTMT decision tree for Versatile Video Coding based on deep neural network

Abdallah, Bouthaina; Belghith, Fatma; Ayed, Mohamed Ali Ben; Nouri, Mohammad

doi:10.1007/s11042-022-13479-7

Cited by 7 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [17], a novel feature based on Statistical Oriented Gradient (SOG) was proposed to extract the feature information of a coding block, and it used SOG to speed up the intra prediction mode decision process. In [18], a fast QTMT partition algorithm based on a CNN-binary tree horizontal (CNN-BTH) network was developed to predict the BTH mode decision at 32 × 32 coding units (CUs). The BTV decision tree algorithm was also predicted at this level by a CNN-binary tree vertical (CNN-BTV).…”

Section: Related Workmentioning

confidence: 99%

A Fast Algorithm for VVC Intra Coding Based on the Most Probable Partition Pattern List

Zhao,

Shang

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

Compared with High-Efficiency Video Coding (HEVC), Versatile Video Coding (VVC) has more flexible division and higher compression efficiency, but it also has higher computational complexity. In order to reduce the coding complexity, a fast algorithm based on the most probable partition pattern list (MPPPL)and pixel content similarity is proposed. Firstly, the MPPPL is constructed by using the average texture complexity difference of the sub-coding unit under different partition modes. Then, the sub-block pixel mean difference is used to decide the best partition mode or shorten the MPPPL. Finally, the selection rules of the reference lines in the intra prediction process are counted and the unnecessary reference lines are skipped by using the pixel content similarity. The experimental results show that compared with VTM-13.0, the proposed algorithm can save 52.26% of the encoding time, and the BDBR (Bjontegarrd delta bit rate) only increases by 1.23%.

show abstract

Section: Related Workmentioning

confidence: 99%

A Fast Algorithm for VVC Intra Coding Based on the Most Probable Partition Pattern List

Zhao,

Shang

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…When the probability is higher than the fixed threshold, the current CU is split directly into two rectangular blocks in the horizontal direction, skipping both the RDcost computation and QT, BTV, TTH, and TTV modes verification. For the BTV algorithm, 21 a new convolution model called CNN-BTV is created according to the three partition levels sized 32 × 32, 16 × 32, and 8 × 32. Its architecture is similar to the CNN-BTH model as shown in Fig.…”

Section: Intra-partition Algorithm Based On Bt Structurementioning

confidence: 99%

“…For the BTH algorithm, 21 a CNN named CNN-BTH is proposed to optimize the intra-BTH splitting decision. In this part, we mainly focus on the three BTHs sized 32×32, 32×16, and 32×8 with BTH depths ranging from 0 to 2.…”

Section: Proposed Intra-qtmt Partition Algorithmmentioning

confidence: 99%

See 1 more Smart Citation

Fast mode decision for CU coding based on CNN for the VVC standard

et al. 2022

Self Cite

View full text Add to dashboard Cite

Versatile Video Coding (VVC) is the latest generation of the video coding standard.In VVC, the advanced quadtree with a nested multitype tree (QTMT) partition structure provides more flexible coding unit (CU) partition sizes compared with the quadtree (QT) decision tree structure applied in the previous High Efficiency Video Coding (HEVC) standard. This flexibility, achieved by the new QTMT partitioning improvement, considerably improves the coding performance while increasing the coding computational complexity caused mainly by the rate distortion optimization processing. To overcome the complexity issue, a fast deep intra QTMT decision tree approach based on a convolution neural network (CNN) is adopted to determine the QTMT depth decision of each 128 × 128 Coding Tree Unit (CTU). The proposed algorithm predicts both the BT depths at 32 × 32 CUs and the QT depths at 64 × 64 using trained CNNs designed for each structure instead of processing the RDcost. Experimental results prove that the suggested deep QTMT approach achieves an important complexity reduction of up to 55.51% compared with the original reference software VTM3.0, with an average of about 35% encoding time reduction accompanied by an insignificant loss in encoding performance.

show abstract