DOC PREVIEW
UT Arlington EE 5359 - AVS-M - data mining algorithm

This preview shows page 1-2-3-19-20-38-39-40 out of 40 pages.

Save
View full document
Premium Document
Do you want full access? Go Premium and unlock all 40 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

By Guided by Term Ramolia Pragnesh R Dr K R Rao Dr Dongil Han Fall 2009 1 Introduction to AVS M Overview of AVS M Complexity present in AVS M encoder Various approaches to reduce complexity Introduction to data mining algorithm C4 5 Project implementation steps AVS M execution and mode and attribute extraction Future work 2 Introduction to AVS M AVS M is the seventh part of video coding standard developed by AVS workgroup of China which aims at mobile applications It has 9 different levels for different formats It supports only progressive video coding hence codes frames only It uses only 4 2 0 chroma sub sampling format It uses only I and P frames 3 Different parts of AVS 10 Part Name 1 System 2 Video 3 Audio 4 Conformance test 5 Reference software 6 Digital media rights management 7 Mobile video 8 Transmit AVS via IP network 9 AVS file format 10 Mobile speech and audio coding Table 1 Different parts of AVS 4 Layered Data Structure G O P Sequence Sequence Picture Slice Picture Macro Block Block Slice Block Macro block 5 AVS M Codec Each MB needs to be intra or inter predicted Switch S0 Fig 1 is used to decide between inter and intra based type of MB Unit size for intra prediction is block size of 4x4 and predictions are derived based on left and upper blocks Inter predictions are derived on blocks of varying sizes 16x16 16x8 8x16 8x8 8x4 4x8 and 4x4 from locally reconstructed frames Transform coefficients are coded by VLC Deblocking filter is applied on reconstructed image 6 Encoder Figure 1 Encoder of AVS M 10 7 Decoder Figure 2 Decoder of AVS M 10 8 Major and Minor tools of AVS M Network abstraction layer NAL Supplemental enhancement information SEI Transform 4x4 integer transform Adaptive quantization of step size varying from 0 63 Intra prediction 9 modes Fig 3 simple 4x4 intra prediction and direct intra prediction Motion compensation 16x16 16x8 8x16 8x8 8x4 4x8 and 4x4 block sizes Quarter pixel interpolation 8 tap horizontal interpolation filter and 4 tap vertical interpolation filter Simplified in loop de blocking filter Entropy coding Error resilience 9 Intra adaptive directional prediction 25 Figure 3 Intra adaptive directional prediction 10 Intra prediction Intra prediction scheme in AVS M brings much simplicity as compared to H 264 baseline profile It uses 4x4 block as the unit for intra prediction It uses 2 modes of prediction in intra prediction intra 4x4 and direct intra prediction Intra 4x4 uses content based most probable intra mode decision as shown in Table 2 to save bits where U and L represents the upper ad left blocks as shown in Fig 4 Upper block U Left block L Current block Fig 4 Current block and neighboring block representation Direct intra prediction brings much of the compression based on trade off decision 11 Intra prediction 1 0 1 2 3 4 5 6 7 8 1 8 8 8 8 8 8 8 8 8 8 0 8 0 0 2 0 0 0 2 0 2 1 8 2 1 2 2 2 2 2 2 2 2 8 2 2 2 2 2 2 2 2 2 3 8 2 1 2 3 4 5 2 7 2 4 8 4 4 2 4 4 4 6 4 4 5 8 5 5 2 5 5 5 6 5 5 6 8 6 6 6 6 6 6 6 6 6 7 8 7 7 2 7 7 7 6 7 7 8 8 0 1 2 3 4 5 6 7 8 U L Table 2 Content based most probable mode decision table 25 Mode 1 is assigned to L or U when the current block does not have Left or Upper block respectively 12 Inter frame prediction Size of the blocks in inter frame prediction can be 16x16 16x8 8x16 8x8 8x4 4x8 and 4x4 depending on the amount of information present within the macro block Motion is predicted up to pixel accuracy If the half pixel mv flag is 1 then it is up to pixel accuracy 8 tap filter F1 1 4 12 41 41 12 4 1 and 4 tap filter F2 1 5 5 1 are used for horizontal and vertical interpolations respectively for pixel MV search and averaging liner interpolation is used for pixel accuracy as shown in Fig 6 13 Inter frame block sizes 7 block sizes are present in AVS M for inter frame prediction 9 4 0 4 1 4 0 8 8 4 4 0 0 0 16 2 3 4 4 8 8 0 1 1 8 8 8 16 0 1 8 8 8 8 4 1 16 8 8 8 0 1 8 2 3 8 16 Figure 5 Inter frame prediction block sizes 14 sub pixel motion estimation by interpolation 16 Figure 6 interpolation of sub pixels hatched lines show half pixels empty circles are quarter pixels and capital letters represent full pixels 15 Error concealment and resilience 3 techniques are used for error concealment forward backward and interactive error concealment For error resilience supplemental enhancement information SEI is sent along with the bit stream which has details of 1 The frame number from which particular block motion starts and 2 Type of motion zooming out in transversal motion in plane etc SEI helps to recover any information lost due to transmission error 16 AVS M encoder complexity variable block sizes in Inter Mode It supports 9 intra 4 4 mode and 1 Direct intra prediction mode Full search for motion estimation gives the optimum result but that comes along with implementation complexity For example assuming FS full search and M block types N reference frames and a search range for each reference frame and block type equal to W we need to examine N x M x 2W 1 2 positions compared to only 2W 1 2 positions for a single reference block type 17 Continued 7 inter prediction modes because of 7 different block sizes 9 intra 4 4 modes and 1 direct intra prediction mode and pixel accuracy in motion vector estimation 18 Various techniques to reduce complexity Intra mode selection algorithm 26 Only intra spatial prediction scheme 27 Fast mode decision algorithm for intra prediction for H 264 AVC 28 Dynamic control of motion estimation search parameters for low complexity H 264 29 Adaptive algorithm for fast motion estimation 30 Data mining algorithm for fast motion estimation 2 19 Data mining algorithm C4 5 Extracts information from data automatically by computational and statistical methods Based on the information extracted develops trees These trees give the decision statement for mode decision Takes decision based on metrics such as MB mean MB variance amplitude of edge detection residual variance etc 20 Goal of this project Implement data mining algorithm c4 5 to decide the inter prediction block mode 21 Implementation steps Select number of frames of a video sequence in QCIF as training sequences Obtain the required attributes off line Encode the training sequence using full complexity AVS M encoder Store the attributes calculated off line and mode decision …


View Full Document

UT Arlington EE 5359 - AVS-M - data mining algorithm

Documents in this Course
JPEG 2000

JPEG 2000

27 pages

MPEG-II

MPEG-II

45 pages

MATLAB

MATLAB

22 pages

AVS China

AVS China

22 pages

Load more
Download AVS-M - data mining algorithm
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view AVS-M - data mining algorithm and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view AVS-M - data mining algorithm and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?