Back to Search Document details
31st Meeting: Geneva, CH, July 2023 2023-07-11 06:55
EE2-4.3: Context modelling for transform coefficients for LFNST/NSPT
Authors: Pavel Nikitin Qualcomm Muhammed Coban Qualcomm Cagan Selim Coban Qualcomm Bappaditya Ray Qualcomm Vadim Seregin Qualcomm Marta Karczewicz Qualcomm
Abstract
This contribution reports test results of EE2-4.3, where modification of the way how the context modelling is done for LFNST/NSPT. The proposed method uses the previous 5 coefficients in the coding order to model the contexts of sig_coeff_flag, coeff_abs_level_greaterX_flag, and for the derivation of Rice parameter for parsing the remainder of the coefficient, rather than using the causal 2D neighborhood for the context derivation.
Meeting Notes
JVET-AE0102 EE2-4.3: Context modelling for transform coefficients for LFNST/NSPT [P. Nikitin, M. Coban, C. S. Coban, B. Ray, V. Seregin, M. Karczewicz (Qualcomm)]
Cited:
JVET-AE0225 X
ARCHIVE :
JVET-AE0102-v2.docx JVET-AE0102-v2.xlsm
Decisions
JVET-AE0102 EE2-4.3: Context modelling for transform coefficients for LFNST/NSPT [P. Nikitin, M. Coban, C. S. Coban, B. Ray, V. Seregin, M. Karczewicz (Qualcomm)]
Citation