Joint Collaborative Team on Video Coding (jct-vc) Contribution

5.2SCE2: Inter-layer texture prediction signalling and deblocking

5.2.1SCE2 summary and general discussion

JCTVC-M0022 SCE2: Summary Report of SHVC Core Experiment on Inter-layer Texture Prediction Signalling and Deblocking [L. Guo, Y. He, D.-K. Kwon, J. Zan, J.-W. Kang]

SHVC Experiment 2 (SCE 2) includes two subtests which are related to inter-layer texture prediction. In subtest 2.1, methods related to inter-layer texture prediction flag signalling were investigated. In subtest 2.2, a method for BS setting of inter-layer texture prediction was evaluated. This report provides a summary of activities and results for this core experiment.

Subtest 2.1: Signalling for inter-layer texture prediction

Test	Changes in Intra-BL flag coding
2.1.1 (TI)	In P and B slice, cu_skip_flag = 1 with an intraBL flag = 1 is used to indicate zero residue for IntraBL, and cu_skip_flag = 1 with intraBL flag = 0 indicates inter skip
2.1.2 (TI)	PU-level signalling Conditionally adding Intra-BL prediction as a candidate in the merge list for non-2Nx2N PU
2.1.3 (TI)	Combination of 2.1.1 and 2.1.2

	Avg. LD-P/RA 2x			Avg. LD-P/RA 1.5x			Avg. LD-P/RA SNR
Test	Y	U	V	Y	U	V	Y	U	V
2.1.1	−0.1%	0.1%	0.1%	−0.2%	0.2%	0.1%	−0.0%	−0.0%	−0.0%
2.1.2	−0.3%	0.0%	0.1%	−0.3%	−0.2%	−0.0%	−0.2%	0.1%	0.3%
2.1.3	−0.4%	0.1%	0.2%	−0.4%	0.0%	0.1%	−0.3%	0.1%	0.3%

	LD-B 2X			LD-B 1.5x			LD-B SNR
Test	Y	U	V	Y	U	V	Y	U	V
2.1.1	−0.1%	0.1%	0.1%	−0.3%	0.3%	0.3%	−0.1%	−0.0%	0.0%
2.1.2	−0.3%	−0.1%	−0.1%	−0.3%	−0.3%	−0.2%	−0.2%	−0.2%	0.1%
2.1.3	−0.4%	0.0%	0.0%	−0.6%	−0.2%	−0.0%	−0.3%	−0.2%	0.1%

The anchor here was IntraBL coding with CU-level selection of inter-layer texture prediction flag.

It was remarked that one difference between the prior proposed IntraBL and RefIdx approaches is that RefIdx operates at the PU level, whereas IntraBL has operated at the CU level. Here, the 2.1.3 approach provides a small gain (0.4%) with some added complexity relative to IntraBL, but in some sense there is a similar gain embedded in the RefIdx approach.

The suggestion is to keep in mind the 0.4% gain opportunity that can strengthen IntraBL when comparing it to the RefIdx approach.

Subtests 2.2: Deblocking for inter-layer texture prediction

Test
2.2 (LG && Qualcomm) JCTVC-M0093	BS =1 for Intra-BL CU Deblocking chroma for both Intra and Intra-BL

	AI 2X			AI 1.5x
Test	Y	U	V	Y	U	V
2.2	−0.1%	−0.2%	−0.2%	−0.1%	−0.3%	−0.1%
	Avg. LD-P/RA 2x			Avg. LD-P/RA 1.5x			Avg. LD-P/RA SNR
Test	Y	U	V	Y	U	V	Y	U	V
2.2	0.0%	−0.2%	−0.2%	−0.1%	−0.3%	−0.4%	0.0%	−0.4%	−0.4%

This has some increase in the complexity of the deblocking filter process.

The potential visual impact would be more important than the PSNR impact.

Does not seem likely to have significant benefit.

5.2.2SCE2 primary contributions

JCTVC-M0216 SCE2: Results of test 1.1 on skipped inter-layer texture prediction (ILTP) block signalling in Inter slice [D.-K. Kwon, M. Budagavi (TI)]
JCTVC-M0217 SCE2: Results of test 1.2 and test 1.3 on inter-layer texture prediction [D.-K. Kwon, M. Budagavi (TI)]
JCTVC-M0093 SCE2.2: Deblocking boundary strength modification for Intra BL [C. Kim, B. Jeon (LG), L. Guo, G. Van Der Auwera, J. Chen, M. Karczewicz (Qualcomm)]

5.2.3SCE2 cross checks

JCTVC-M0341 SCE2: Cross-check for test 1.1 on skipped inter-layer texture prediction (ILTP) block signalling in Inter slice [E. Alshina] [late]
JCTVC-M0054 Cross-check for SCE2: Results of test 1.2 and test 1.3 on inter-layer texture prediction (JCTVC-M0217) [Z. Ma, F. Fernandes] [late]
JCTVC-M0174 SCE2 Cross-check of CE2 Sub-test 2 [A. Norkin (Ericsson)] [late]
JCTVC-M0399 SCE1: Cross-check of SCE1 2.1 on gradient based intra prediction [K. Rapaka (Qualcomm)] [late]