Home

  Year of Polar Prediction – from research to improved environmental safety | World Meteorological Org

It is therefore expected that advances in polar prediction will lead to improved weather forecasts and climate predictions both for polar regions and densely populated countries. (...) The science and technology improvements developed under the umbrella of the Year of Polar Prediction will both strengthen physical understanding of the Arctic and Antarctic, and provide the foundation of better observational and predictive systems for the future. Notes to Editors: A new animation on the Year of Polar Prediction is available at https://www.youtube.com/watch?

Score: 681879.2 - https://public.wmo.int/en/medi...-improved-environmental-safety

  Joint Collaborative Team on Video Coding (JCT-VC) Contribution

Despite the improvements on intra prediction scanning, very few work have been done on finding a more efficient scan order for inter-frame prediction where a large portion of data are compressed with. (...) The algorithm has three steps: 1) For each inter predicted frame, initialize a set of 2-D arrays {S1[n][n], S2[n][n], …, Sm[n][n]}, where m is the number of inter prediction modes, n is the size of the transform in each dimension. (...) More results will be provided. 5 Conclusion We propose a prediction-based adaptive transform coefficient scanning method for inter prediction coding in H.264.

Score: 681849.17 - https://www.itu.int/wftp3/av-a...04_A_Dresden/JCTVC-A020_r1.doc

  Joint Video Team (MPEG+ITU) Document

Functional stage : Constrained Intra prediction Purpose : Check that decoder handles constrained intra prediction. (...) Both temporal and spatial direct prediction is used for direct prediction. Stored B slice is included in the stream. (...) Both temporal and spatial direct prediction is used for direct prediction. Memory management control operation is used.

Score: 681729.64 - https://www.itu.int/wftp3/av-a...003_09_SanDiego/JVT-I045d0.doc

  Joint Video Team (MPEG+ITU) Document

Functional stage : Constrained Intra prediction Purpose : Check that decoder handles constrained intra prediction. (...) Both temporal and spatial direct prediction is used for direct prediction. Stored B slice is included in the stream. (...) Both temporal and spatial direct prediction is used for direct prediction. Memory management control operation is used.

Score: 681729.64 - https://www.itu.int/wftp3/av-a.../2003_09_SanDiego/JVT-I045.doc

  Joint Video Team (MPEG+ITU) Document

The mechanism of proposed alternative prediction is basically same as that of bi-predictive prediction. (...) Neighboring blocks for alternative prediction Table 1. Adaptive direct mode Selection of prediction for direct mode forward MV in collocated block relative to index 0 of current picture others direct_reference_usable ‘1’ direct 4x4 bi-predictive ‘0’ 4x4 bi-predictive The proposed description for alternative prediction, named 4x4 bi-predictive, is provided below. 4x4 bi-predictive If the collocated block does not use forward prediction picture relative to index 0 of forward reference picture set, or direct_reference_usable in picture_laye_rbsp() or slice_header() is “0”, the 4x4 bi-predictive prediction is applied instead of direct mode ( this is the case the solution 1 of section 2 is applied ). (...) Note that if both 8x8 block immediately above and to the left has have same direction prediction, e.g. forward or backward, that prediction is used instead of bi-predictive.

Score: 680930 - https://www.itu.int/wftp3/av-a...002_07_Klagenfurt/JVT-D050.doc

  Joint Collaborative

However, intra prediction of chroma components in YUV 4:2:0 videos uses a limited set of possible predictions available for coding of luma components. (...) When the YUV colour-space is used, each of the Y, U and V colour planes are independently predicted and transform coded. While the Y component can have up to 34 intra prediction directions [8], the U and V component share the same intra prediction mode, and have only 5 possible intra prediction modes: DC, vertical, horizontal, diagonal down-right (or plane) and the signalled luma intra prediction mode [1]. (...) In our implementation, CLTM is available as another chroma intra prediction mode, in addition to DC, horizontal, vertical and diagonal down-right (or plane) prediction.

Score: 680687.16 - https://www.itu.int/wftp3/av-a..._10_C_Guangzhou/JCTVC-C040.doc

  Core Experiment on separate prediction modes for 4:4:4 coding (CE-8)

JVT-Q032 Introduction of improved entropy coding of intra prediction modes is proposed in this contribution. In some proposals, independent intra prediction for each color component is adopted, which can improve prediction efficiency but leads increase of overhead bits to code additional intra prediction modes. The efficiency of entropy coding for intra prediction modes is investigated for the case that independent intra prediction per each color component is employed. (...) Murakami, “An improved entropy coding for intra prediction modes” [4] JVT-Q033 S. Sekiguci, Y. Isu, Y.

Score: 680660.4 - https://www.itu.int/wftp3/av-a...site/2005_10_Nice/JVT-Q309.doc

  untitled

untitled 1 AVC based intra prediction for AVC based intra prediction for improved visual qualityimproved visual quality A.A.Tabatabai, T. (...) Suzuki SonySony JCTVC-A030 Overview • Additional new tool to JCTVC-A103 • Motivation: – Current AVC intra prediction has coarsely quantized prediction direction (i.e., 8) – More precision in the prediction direction results in better prediction • However, the bit overhead for signaling directions may be significant • Proposal: – Derive a neighbor's suggested direction using edge detection • Edge in natural images are commonly continuous – Allow the suggested direction to be adjusted by the encoder • Account curved edge EdgeEdge--Based Intra Prediction (EBIP)Based Intra Prediction (EBIP) • Get edge vectors from neighbors together with a confidence measure: – In a neighborhood, perform edge detection for each pixel using for example Sobel operators; – Discard edge vectors that do not cross the block boundary; • Can be easily done with a lookup table. – Discard edge vectors that have small norms. 3 neighbors’ suggested prediction direction *θ discarded edge vectors ia v current block min slopemax slo pe EBIP (continued)EBIP (continued) 4 • Direction adjustment: – Based on the confidence measure s we introduce 2m+1 new modes: – The following signaling is used for the 2m+10 modes 0 to 8 mode 9 to 2m+9 Mo st pro ba ble rest 1 0XXX 1 0 Writing to bit stream * 1 0 * 2 if ( ) otherwise k s s k θ δθ θ δθ ⎧ ⎫+ < ⎨ ⎬ +⎩ ⎭ 0, 1, 2,...k = ± ± 2 1δθ δθ< k Symb 0 1 1 000 -1 001 2 010 -2 011AVC Tr un ca te d un ar y co de s Experiment • Implemented in the KTA 2.6r1 software for 4x4 and 8x8 blocks – although this method can be directly applied to arbitrary block sizes • All frames (entire sequence) are intra- coded • 5 different QPs (23, 27, 31, 35, 39) – Low BD Rate (largest 4 QPs) – High (smallest 4 QPs) Results Sequence/Resolution Low Bitrate % High Bitrate % CIF Foreman (300 frames) -5.85 -5.18 CIF average -5.85 -5.18 Basketball Pass -3.13 -3.27 Blowing Bubbles -2.70 -2.66 Race Horse -3.54 -3.34 Class D BQ Square -3.28 -3.30 Class D average -3.16 -3.14 Basketball Drill -9.38 -9.97 BQ Mall -3.06 -3.18 Party Scene -1.56 -1.54 Class C Race Horse -3.28 -2.87 Class C average -4.32 -4.39 BQ Terrace -5.83 -4.45 Park Scene -0.75 -0.78 Kimono -0.95 -1.26 Basketball Drive -4.86 -4.63 Class B Cactus -4.26 -3.98 Class B average -3.33 -3.02 AVERAGE -3.75 -3.60 Summary • Edge-based intra prediction is proposed – Additional modes are added to AVC intra prediction – The gain is up to -10% – The average gain is -3.75% (low), -3.60% (high) • Propose to study by CE

Score: 680474.66 - https://www.itu.int/wftp3/av-a...10_04_A_Dresden/JCTVC-A030.pdf

  Joint Video Team (MPEG+ITU) Document

The chroma in intra macroblocks is predicted in a manner very similar to the luma block in Intra_16x16 macroblock type (subclause 8.3.2), using one of four prediction modes. The same prediction mode is applied to both chroma blocks, but it is independent of the prediction mode used for the luma. (...) Let Pred(x,y) with x,y = 0..7 denote the prediction for the 8x8 chroma block samples. There are four different prediction modes as specified in subclauses 8.3.3.1 to 8.3.3.4.

Score: 680213.5 - https://www.itu.int/wftp3/av-a...ite/2002_12_Awaji/JVT-F026.doc

  Joint Video Team (MPEG+ITU) Document

The chroma in intra macroblocks is predicted in a manner very similar to the luma block in Intra_16x16 macroblock type (subclause 8.3.2), using one of four prediction modes. The same prediction mode is applied to both chroma blocks, but it is independent of the prediction mode used for the luma. (...) If none of the neighbouring samples are available, Mode 2 (DC prediction) is the only possible 8x8 chroma prediction mode. 8.3.3.1 Mode 0: DC prediction process The 8x8 block of chroma samples is predicted in four separate 4x4 sub-blocks, labeled A-D as shown in Figure 8-7.

Score: 680213.5 - https://www.itu.int/wftp3/av-a...e/2002_12_Awaji/JVT-F026r1.doc