Home

  Joint Video Team (MPEG+ITU) Document

(See Fig.1~4) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% FootballBusForemanMobile ctxIdxInc2,flagOff ctxIdxInc1,flagOff ctxIdxInc0,flagOff ctxIdxInc2,flagOn ctxIdxInc1,flagOn ctxIdxInc0,flagOn Fig.1 Probability distribution at each context 0% 20% 40% 60% 80% 100% FootballBusForemanMobile ctxIdxInc0,flagOff ctxIdxInc0,flagOn Fig.2 Probability distribution when ctxIdxInc = 0 0% 20% 40% 60% 80% 100% FootballBusForemanMobile ctxIdxInc1,flagOff ctxIdxInc1,flagOn Fig.3 Probability distribution when ctxIdxInc = 1 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% FootballBusForemanMobile ctxIdxInc2,flagOff ctxIdxInc2,flagOn Fig.4 Probability distribution when ctxIdxInc = 2 Fig.1 gives the whole distribution of three contexts. (...) If the base macroblock is coded as an intra mode, the probability of the base_mode_flag to be set increases up to 90% (refer to Fig.6) and in the opposite case that the base macroblock is coded as inter mode, the probability of the base_mode_flag to be set is about 55% (refer to Fig.7). The following figures show that the probability of base_mode_flag depends on MB type of base macroblock. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% BusFootballForemanMobile flagOff,baseInter flagOff,baseIntra flagOn,baseInter flagOn,baseIntra Fig.5 Probability distribution according to MB type of base macroblock 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% BusFootballForemanMobile flagOff,baseIntra flagOn,baseIntra Fig.6 Probability distribution when MB type of base macroblock is intra 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% BusFootballForemanMobile flagOff,baseInter flagOn,baseInter Fig.7 Probability distribution when MB type of base macroblock is inter To clarify the effect of MB type of base macroblock on the base_mode_flag, we have analyzed the probability distribution depending on MB type of base macroblock and the current context model using Palma test condition. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% FootballBusForemanMobile ctxIdxInc2,flagOff,baseInter ctxIdxInc2,flagOff,baseIntra ctxIdxInc1,flagOff,baseInter ctxIdxInc1,flagOff,baseIntra ctxIdxInc0,flagOff,baseInter ctxIdxInc0,flagOff,baseIntra ctxIdxInc2,flagOn,baseInter ctxIdxInc2,flagOn,baseIntra ctxIdxInc1,flagOn,baseInter ctxIdxInc1,flagOn,baseIntra ctxIdxInc0,flagOn,baseInter ctxIdxInc0,flagOn,baseIntra Fig.8 Probability distribution according to MB type of base macroblock and context model 0% 20% 40% 60% 80% 100% FootballBusForemanMobile ctxIdxInc0,flagOff,baseIntra ctxIdxInc0,flagOn,baseIntra Fig.9 Probability distribution when base macroblock is intra and ctxIdxInc = 0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% FootballBusForemanMobile ctxIdxInc0,flagOff,baseInter ctxIdxInc0,flagOn,baseInter Fig.10 Probability distribution when base macroblock is inter and ctxIdxInc = 0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% FootballBusForemanMobile ctxIdxInc1,flagOff,baseIntra ctxIdxInc1,flagOn,baseIntra Fig.11 Probability distribution when base macroblock is intra and ctxIdxInc = 1 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% FootballBusForemanMobile ctxIdxInc1,flagOff,baseInter ctxIdxInc1,flagOn,baseInter Fig.12 Probability distribution when base macroblock is inter and ctxIdxInc = 1 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1234 ctxIdxInc2,flagOff,baseIntra ctxIdxInc2,flagOn,baseIntra Fig.13 Probability distribution when base macroblock is intra and ctxIdxInc = 2 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1234 ctxIdxInc2,flagOff,baseInter ctxIdxInc2,flagOn,baseInter Fig.14 Probability distribution when base macroblock is inter and ctxIdxInc = 2 As shown figures, if MB type of base macroblock is inter, the probability of base_mode_flag is very similar to that of JSVM 2.0.

Score: 730622.85 - https://www.itu.int/wftp3/av-a...te/2005_07_Poznan/JVT-P104.doc

  North Eurasian Climate Outlook Forum (NEACOF-21), Moscow | World Meteorological Organization

Warmer than usually with a probability of 45-60% may be in western half of Belarus, western Ukraine and southern Moldova. (...) In the rest of Central Asia the temperature above normal is expected with a 45-60% probability.                                       Figure 1. (...) An excessive precipitation is predicted (with probability up to 50 %) in the central and eastern regions of Yakutia and in the Far East region.

Score: 729921.04 - https://public.wmo.int/en/medi...outlook-forum-neacof-21-moscow

  IBD Estimation in Pedigrees

Kohler Population Studies Center University of Pennsylvania Overall Structure of the Survey on Aging in SSA 2 major components: • Household Interview • Individual Interview • Both will be linked based on HH ID and individual ID Overall Sampling Design  Nationally representative stratified random sample of households that include at least 1 household member age 60 years and older  Household sample surveys:  Key source for data on social phenomena  Are among the most flexible methods of data collections  In theory almost any population-based subject can be investigated through household surveys  Only probability samples following well-established sampling procedures are suitable for making inferences from the sample population to the larger population that it is designed to represent  Snow-ball or convenience samples are not suitable for this survey Overall Sampling Design cont’d  Probability sampling in the context of household surveys:  Refers to the means by which elements of the target population are selected for inclusion in the survey  In order to be cost-effective, most household surveys are not implemented as simple random samples  Sampling procedure usually includes stratification to ensure that the selected sample actually is spread over geographic sub-areas and population subgroups  This sampling design usually uses clusters of households in order to keep costs to manageable level General Principals of the Survey on Aging in SSA Target population: individuals age 60+ and older Household sample: Nationally representative clustered random sample of households that include household members age 60+ yrs. Selection of household members: All regular household members age 60+ in the sampled household and their spouses if these are age-eligible and co-resident General Principals of the Survey on Aging in SSA Use of an existing sampling frame: clustered random sample of households can only be obtained from existing sampling frame which is a complete list of statistical units covering the target population  Census frame, complete list of villages/communities or sampling list from other nationally representative surveys  Sampling frame: is a complete list of sampling units that entirely covers the target population  Conventional sampling frame: list of enumeration areas (EA) from a recently completed census  EA: geographic area which usually groups a number of households together for convenient counting purposes General Principals of the Survey on Aging in SSA Stratification: process in which the sample is designed into sub- groups or strata that are as homogeneous as possible;  Within each stratum the sample is designed and selected independently; Two-stage cluster sampling procedure:  Cluster: a group of adjacent households which serves as the primary sampling unit (PSU) General Principals of the Survey on Aging in SSA Full coverage of the target population: should be nationally representative and cover 100% of the target population; that is no subpopulations age 60+ are systematically excluded; Probability sampling: sample should be obtained as probabilistic sample based on existing sampling frame using established sampling procedures;  Only way to obtain unbiased estimation and to be able to evaluate the sampling errors  Excluded are purposive sampling, quota sampling, and other uncontrolled non-probability methods because they cannot provide evaluation of precision and confidence of survey findings General Principals of the Survey on Aging in SSA Full coverage of the target population: should be nationally representative and cover 100% of the target population; that is no subpopulations age 60+ are systematically excluded; Probability sampling: sample should be obtained as probabilistic sample based on existing sampling frame using established sampling procedures;  Only way to obtain unbiased estimation and to be able to evaluate the sampling errors  Excluded are purposive sampling, quota sampling, and other uncontrolled non-probability methods because they cannot provide evaluation of precision and confidence of survey findings Sample Size  Sample size must take into account competing needs so that costs and precisions are optimally balanced  Sample size must also address the needs of users who desire for sub-populations of sub-areas domains  Sample size is determined by the trade-offs between survey precision, data quality, organizational capacities and survey budget;  In the case of Malawi this is about 2,000 respondents (men and women) Conducting a household listing and pre-selection of households  Data quality is enhances if eligible households are preselected for participating in the study  In many SSA countries recent and reliable household listings in EAs that carefully enumerates older individuals is not available  Hence, we suggest to conduct a specific household listing in selected EAs that provides a well-grounded basis for selecting respondents  Interviewers than interview only pre-selected eligible households  STEPS:  Household listing operation conducted before the survey  Pre-selection of households from this list  Selected Households are interviewed Overall Sampling Design cont’d  Two stage sample design is well-established approach for implementing household surveys  1st stage: select a sample of EAs with probability proportional to size (PPS);  Within each stratum a sample of predetermined number of EAs is selected independently with probability proportional to size, where size is measured in terms of older individuals age 60+;  If size of pop age 60+ is not available, and variations in age structures are relatively modest, then total pop size can be used  All households in the EAs are listed  2nd stage: after complete listing in EAs, a fixed number of households with individuals age 60+ is selected by equal probability sampling in the EAs Interviewing all individuals age 60+ in the HH Advantages:  Maximize the number of respondents for a given sample of HH  Cost effective to achieve the sample size  Analytical advantages so that interactions among spouses, within and between household variation of outcomes can be investigated Disadvantages:  Lower statistical power given the within household correlation of observations  Logistical challenges in the fieldwork Sample Take per Cluster  How many eligible individuals to interview per EA  DHS recommends 25-30 individuals  Because there will be more than 1 age-eligible individual per household, less than 24-30 households per PA need to be selected  If a sampled HH has 1.5 age-eligible individuals on average, than a sample take per cluster of 25-30 individuals results in the selection of 17-20 households per cluster  With 2,000 individuals sample size: 67-80 clusters have to be selected  If sample is stratified, these considerations should be conducted stratum- specific Sample Take per Cluster  This fixed sample take per cluster is:  Easy for survey management and implementation  But requires sampling weights that vary within clusters

Score: 725909.86 - https://www.un.org/en/developm...15/ikohler-sampling-design.pdf

  WHO | Ebola virus disease update - west Africa

عربي 中文 English Français Русский Español Home Health Topics All topics » A B C D E F G H I J K L M N O P Q R S T U V W X Y Z Resources » Fact sheets Facts in pictures Multimedia Publications Questions & answers Tools and toolkits Popular » Coronavirus disease (COVID-19) Ebola virus disease Air pollution Hepatitis Top 10 causes of death World Health Assembly » Countries All countries » A B C D E F G H I J K L M N O P Q R S T U V W X Y Z Regions » Africa Americas South-East Asia Europe Eastern Mediterranean Western Pacific WHO in countries » Statistics Cooperation strategies Democratic Republic of the Congo »   Newsroom All news » News releases Statements Campaigns Commentaries Events Feature stories Speeches Spotlights Newsletters Photo library Media distribution list Headlines » Timeline: WHO's COVID-19 response »   Emergencies Focus on » COVID-19 pandemic Ebola virus disease outbreak DRC 2021 Syria crisis Crisis in Northern Ethiopia Afghanistan Crisis Latest » Disease Outbreak News Travel advice Situation reports Weekly Epidemiological Record WHO in emergencies » Surveillance Research Funding Partners Operations Independent Oversight and Advisory Committee Coronavirus disease outbreak (COVID-19) » WHO © Credits Data Data at WHO » Global Health Estimates Health SDGs Mortality Triple billion targets Data collections Dashboards » COVID-19 Dashboard Triple Billion Dashboard Health Equity monitor Mortality Highlights » GHO SCORE Insights and visualizations Data collection tools Reports World Health Statistics 2021 » WHO © Credits About WHO About WHO » People Teams Structure Partnerships Collaborating Centres Networks, committees and advisory groups Transformation Contact us » Governance » World Health Assembly Executive Board Election of Director-General Governing Bodies website Our Work » General Programme of Work WHO Academy Activities Initiatives Better health for everyone » WHO © Credits Skip to main content Access Home Alt+0 Navigation Alt+1 Content Alt+2 Emergencies preparedness, response Menu Home Alert and response operations Diseases Biorisk reduction Disease outbreak news Ebola virus disease update - west Africa Disease outbreak news 28 August 2014 Epidemiology and surveillance The total number of probable and confirmed cases in the current outbreak of Ebola virus disease (EVD) in the four affected countries as reported by the respective Ministries of Health of Guinea, Liberia, Nigeria, and Sierra Leone is 3069, with 1552 deaths. (...) The distribution and classification of the cases are as follows: Guinea, 647 cases (482 confirmed, 141 probable, and 25 suspected), including 430 deaths; Liberia, 1378 cases (322 confirmed, 674 probable, and 382 suspected), including 694 deaths; Nigeria, 17 cases (13 confirmed, 1 probable, and 3 suspected), including 6 deaths; and Sierra Leone, 1026 cases (935 confirmed, 37 probable, and 54 suspected), including 422 deaths. Confirmed, probable, and suspect cases and deaths from Ebola virus disease in Guinea, Liberia, Nigeria, and Sierra Leone Confirmed Probable Suspect Totals Guinea Cases 482 141 25 648 Deaths 287 141 2 430 Liberia Cases 322 674 382 1 378 Deaths 225 301 168 694 Nigeria Cases 13 1 3 17 Deaths 5 1 0 6 Sierra Leone Cases 935 37 54 1 026 Deaths 380 34 8 422 Totals Cases 1 752 853 464 3 069 Deaths 897 477 178 1 552 Note: Cases are classified as confirmed (any suspected or probable cases with a positive laboratory result); probable (any suspected case evaluated by a clinician, or any deceased suspected case having an epidemiological link with a confirmed case where it has not been possible to collect specimens for laboratory confirmation); or suspected (any person, alive or dead, suffering or having suffered from sudden onset of high fever and having had contact with: a suspected, probable or confirmed Ebola case, or a dead or sick animal; or any person with sudden onset of high fever and at least three of the following symptoms: headache, vomiting, anorexia/loss of appetite, diarrhoea, lethargy, stomach pain, aching muscles or joints, difficulty swallowing, breathing difficulties, or hiccup; or any person with unexplained bleeding; or any sudden, unexplained death).

Score: 721894.5 - https://www.who.int/csr/don/2014_08_28_ebola/en/

  Step-by-Step Guide to the Estimation of Child Mortality

The basic relationship between the proportion of chil- dren dead by age group of mother and the probability of dying in childhood can be illustrated by a very simple example. (...) Since in a life table the probability of dying by age 4, q(4), must be greater than the probability of dying by age 2, q(2), a given proportion of children dead for women aged 22 would indicate lower mortality risks in an early-fertility population than in a late-fertility popula- tion. (...) CORRESPONDENCE BETWEEN OBSERVED PROPORTIONS OF CHIL- DREN DEAD BY AGE GROUP OF MOTHER AND ESTIMATED PROBABILITIES OF DYING ESTIMATING TIME TRENDS OF MORTALITY The method originally developed by Brass assumed that mortality was constant, so that cohort and period probabilities of dying were identical.

Score: 721770.9 - https://www.un.org/en/developm...s/estimate/childmort/chap3.pdf

  DEFICIENCY (HAZARD)

Revised risk assessment if only this solution is to be implemented: 10.Probability: $ DDDD 11. Severity: 12. Level of risk: 13. (...) Revised risk assessment if only this solution is to be implemented: 18.Probability: $ XXXX 19. Severity: 20. Level of risk: 21. (...) Revised risk assessment if only this solution is to be implemented: 26.Probability: $ 27. Severity: 28. Level of risk: 29.

Score: 719234 - https://www.icao.int/NACC/Docu.../ANDEFWS/ANDeficienciesEx4.doc

  The forward integer transform used in the current H

. · Table-based interval division similar to the M-coder · Scaled probability estimation similar to the M-coder · Byte-based output in the encoder and byte-based input in the decoder instead the bit-based input and output in the M-coder · Probability estimation Totally 40 LPS probability states are used in the coder and they are realize as a markov chain, which allows a table-driven implementation of probability estimation. We recursively define the following set of representative LPS probability states { } 39 38 2 1 0 , , , , p p p p p P L : 39 , 38 , 3 , 2 , 1 5 . 0 5 . 0 1 8 0 L = ´ = = - i p p p i i So we have 5 . 0 8 ´ = + i i p p . (...) We have the following relationships: ( ) ( ) ( ) ( ) ( ) ( ) 1 3 5 . 0 7 & 7 & 7 & 3 - ³ ´ ³ >> - ´ = ´ ´ ´ = ´ >> I MI p I MI I MI i p I MI p I MI p I p I i i i i i i The first property allows us to use much less table storage for the interval subdivision, the second and the third relationships allow us to estimate the position of the most significant bit of the current interval without loop. State Probability Next_State_LPS Next_State_MPS Switch_MPS 0 0.50000 0 1 1 1 0.45850 0 2 0 2 0.42045 1 3 0 3 0.38555 2 4 0 4 0.35355 2 5 0 5 0.32421 3 6 0 6 0.29730 4 7 0 7 0.27263 5 8 0 8 0.25000 5 9 0 9 0.22925 6 10 0 10 0.21022 7 11 0 11 0.19278 8 12 0 12 0.17678 8 13 0 13 0.16210 9 14 0 14 0.14865 10 15 0 15 0.13631 10 16 0 16 0.12500 11 17 0 17 0.11463 11 18 0 18 0.10511 12 19 0 19 0.09639 12 20 0 20 0.08839 13 21 0 21 0.08105 13 22 0 22 0.07433 14 23 0 23 0.06816 14 24 0 24 0.06250 15 25 0 25 0.05731 15 26 0 26 0.05256 15 27 0 27 0.04819 16 28 0 28 0.04419 16 29 0 29 0.04053 16 30 0 30 0.03716 17 31 0 31 0.03408 17 32 0 32 0.03125 17 33 0 33 0.02866 18 34 0 34 0.02628 18 35 0 35 0.02410 18 36 0 36 0.02210 18 37 0 37 0.02026 18 38 0 38 0.01858 19 39 0 39 0.01704 19 39 0 Table 1: Probability states and its transition rules for probability estimation 2 Table-based interval division First the interval I is mapped to a quantized value Q, shown as following: ( ) ( ) ( ) 15 & 4 - >> = I MI I Q . then the updated I is gotten as following (s denotes the current LPS state): ( ) ( ) ( ) ( ) ( ) R I I else R I LPS if s I MI s Q RTAB R - = = >> >> - << = 3 8 7 & , and update the MI(I) lastly: ( ) ( ) ( ) ( ) - - == << + >> = - ) ( 0 ) ( 1 & 1 3 I MI I MI I if s I MI LPS if Q State 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0 132 140 148 156 164 172 180 188 196 204 212 220 228 236 244 252 1 121 128 136 143 150 158 165 172 180 187 194 202 209 216 224 231 2 111 118 124 131 138 145 151 158 165 172 178 185 192 198 205 212 3 102 108 114 120 126 133 139 145 151 157 163 170 176 182 188 194 4 93 99 105 110 116 122 127 133 139 144 150 156 161 167 173 178 5 86 91 96 101 106 112 117 122 127 132 137 143 148 153 158 163 6 78 83 88 93 98 102 107 112 117 121 126 131 136 140 145 150 7 72 76 81 85 89 94 98 103 107 111 116 120 124 129 133 137 Table 2: RTAB(Q, state) only states from 0…7 is stored because others can be gotten by shift 3 Programs Data Structure of the encoder: Unsigned short low=0, range=0xffff; Int pos=7; //MI(range)-8 Int outstanding=0; //used in the renormalization Unsigned char store; //store the byte Encoder: val denotes the input, Pseq denotes the current LPS state Int temp=(Pseq>>3); Int sel=(Pseq&7); Int rLPS=(RTAB[sel][(range>>(pos+4))&15]<<(pos-temp)); Int sLow; If( val== LPS ) { Range=rLPS; pos-=(temp+1); if((range>>pos)<256) pos--; } Else { Range-=rLPS; sLow=low+rLPS; if(sLow>pos)<256) pos--; } If(range<256) { If( ( (low&255)+range ) > 256 ) { If((low>>8)==255 ) Outstanding++; Else { If(outstanding) { outByte(store); outByte(0xff, outstanding-1); } outstanding=1; Store=low>>8; } } Else { If(outstanding) { outByte(store); outByte(0xff, outstanding-1); } outByte(low>>8); outstanding=0; } Low<<=8; Range<<=8; } Update the probability estimation Data structure of the decoder: Unsigned short sub=readByte(2); //read two bytes from the bitstream Unsigned short range=0xffff; Int pos=7; //MI(range)-8 Decoder: val denotes the output, Pseq denotes the current LPS states Int temp=(Pseq>>3); Int sel=(Pseq&7); Int rLPS=(RTAB[sel][(range>>(pos+4))&15]<<(pos-temp)); If( rLPS>sub ) { val=LPS; range=rLPS; pos-=(temp+1); if((range>>pos)<256) pos--; } Else { val=1-LPS; range-=rLPS; sub-=rLPS; If((range>>pos)<256) pos--; } If(range<256) { sub<<=8; range<<=8; sub+=ReadByte(1); //read a byte as the low byte of sub } Update the probability estimation Reference [1].

Score: 718361.3 - https://www.itu.int/wftp3/av-a...eo-site/0601_Ban/VCEG-AB08.doc

  ITU - Telecommunication Standardization Sector

Context modeling provides estimates of conditional probabilities of the coding symbols. Utilizing suitable context models, given inter-symbol redundancy can be exploited by switching between different probability models according to already coded symbols in the neighborhood of the current symbol to encode. 2. (...) This is extremely beneficial for symbol probabilities much greater than 0.5, which often occur with efficient context modeling. (...) The provided models serve as a probability estimation of the related bins. After encoding of each bin, the related model will be updated with the encoded binary symbol.

Score: 718018.36 - https://www.itu.int/wftp3/av-a...eo-site/0105_Por/HHI-CABAC.doc

  \(03/09/2010\) Veronique Arthaud - Chapter 5 Revised

Also, the time required to complete the exercise is shorter. 5.6 In probability sampling, the units are selected in such a way that each unit (an outlet or a product) has a known non-zero probability of selection. (...) We then go on to consider some non- probability techniques. Reasons for using non-probability sampling 5.28 No sampling frame is available. (...) So the absence of sampling frames is not a good enough excuse for not applying probability sampling. 5.29 Bias resulting from non-probability sampling is negligible.

Score: 717145.1 - https://www.ilo.org/public/eng...d/cpi/corrections/chapter5.pdf

  ITU-T Video Coding Experts Group Document

One symbol does not dominate the probability and the UVLC code forms a poor model of the distribution near the origin. (...) The merged code is able to match the probability distribution both near the origin and asymptotically. (...) At low quantizers the performance of the UVLC coder is limited by the distribution of codeword probabilities. To achieve coding near the entropy limit in these cases, the structure of the code must be modified to enable it to more nearly match the probability distribution of symbols.

Score: 717048.9 - https://www.itu.int/wftp3/av-a...video-site/0008_Por/q15k45.doc