Research areas



Yüklə 193,76 Kb.
səhifə3/3
tarix03.11.2017
ölçüsü193,76 Kb.
#28861
1   2   3

2.2 Technical requirements





  • Low/Scalable complexity and power consumption

  • High compression efficiency in high quality imaging

  • Flexibility of implementation (in terms of required resources)

  • Compressed domain manipulation

  • Efficient 3D region of interest decoding/access

  • Unique algorithm core for compression of multiple dimensions images (spatial, components, time, etc.)

Several research projects can be explored related to AIC. Pl access



http://www.bilsen.com/index.htm?http://www.bilsen.com/aic/

AIC Advanced Image Coding (Beyond JPEG & JPEG 2000) ISO/IEC standardization process


I have several files on AIC. I frame only image coding in H.264 versus JPEG-2000 performance comparison.

F. Wu, C. Lan and G. Shi, “Compress Compound Images in H.264/MPEG-4 AVC by Fully Exploiting Spatial Correlation” ISCAS2009 May, Taipei, Taiwan. Paper ID 1109, track 16.1 


JPEG2000:
D. T. Lee, “ JPEG 2000: Retrospective and new developments”, Proc. IEEE vol. 93, pp.32-41, Jan. 2005. (MANY VALUABLE REFERENCES)
This paper describes four new parts designed for important new applications
Part 9 JPIP Interactivity tools, application programming interface and protocols,

Part 8 JPSEC Secure JPEG 2000

Part 10 JP3D 3-D and floating point data

Part 11 JPWL Wireless
These new parts can be the basis for many research projects.
W.D Neve et al, “ Improved BSDL – based content adaptation for JPEG2000 and HD photo (JPEG XR)”, SP:IC, vol. 24,, issue 6, pp. 452-467, July 2009. Good paper for research/projects.
DIGITAL VIDEO CONSORTIUM (DVC) has developed a vision optimized (visual discrimination model – VDM) MPEG-2 Encoder/Decoder which is superior to MPEG-2 (file DVCencoder has PP slides- Also Ch.12 HVS based perceptual video coders authors – A. Pica, M. Isnardi and J. Lubin in the handbook edited by H.R. Wu and K.R. Rao, Digital video image quality and perceptual coding – Taylor and Francis, 2006.) I have the related file DVCencoder which is developed by Sarnoff Corporation.. EXTEND THIS CONCEPT TO H.264/MPEG-4 PART 10 ENCODERS – ALL LEVELS AND PROFILES. Evaluate and compare the performances of H.264 with and without the VDM. To implement this software licensing agreement with Sarnoff Corporation is required. Jennie Abraham is working on this (doctoral dissertation).
JVT-V204 "New profiles for professional applications" amendment to ITU-T Rec. H.264 & ISO/IEC 14496-10 (Amendment 2 to 2005 edition) 2/13/2007 research/projects based on these new profiles. Details are as follows: (file JVT-V204) www.itu.int/rec/T-REC-H.264 (4/10/07)
This document is a draft amendment to ITU-T Rec. H.264 & ISO/IEC 14496-10 creating a set of new profiles intended primarily for professional applications. It also defines two new types of supplemental enhancement information (SEI) messages.

One such new profile is the High 4:4:4 Predictive profile. The High 4:4:4 Predictive profile, as drafted herein, has two different 4:4:4 operation modes depending upon the value of a new syntax element, separate_ color_plane_flag that is present in the sequence parameter set. When separate_

color_plane_flag is equal to 0, each macroblock contains both luma and chroma samples, and a decoding process similar to the luma decoding process that is used in the other profiles is used to decode the luma and chroma samples in each such macroblock. When separate_

color_plane_flag is equal to 1, the decoding process for monochrome


each color plane individually as a distinct picture. In addition, a new intra decoding process that can be used by encoders to enable relatively-efficient lossless coding is also added for use when the qpprime_y_zero_transform_bypass_flag syntax element is equal to 1 and QP'Y is equal to 0. In the new High 4:4:4 Predictive profile, the bit depth is also extended up to 14 bits per sample.

In addition to adding the definition of the High 4:4:4 Predictive profile, four other profiles are also specified in this amendment. These profiles, referred to as the High 10 Intra, High 4:2:2 Intra, High 4:4:4 Intra, and CAVLC 4:4:4 Intra profiles, serve to enable applications demanding simple random-access and editing applications with low delay capability. Each of these profiles contains coding capabilities that similar to those of another corresponding profile, except for elimination of support for the decoding processes that involve inter-picture prediction and, in the case of the CAVLC 4:4:4 Intra profile, the additional elimination of support for the CABAC parsing process.

The two added SEI messages are the post-filter hint SEI message and the tone mapping information SEI message. The post-filter hint SEI message provides the coefficients of a post-filter or correlation information for the design of a post-filter for potential use in post-processing of the output decoded pictures to obtain improved displayed quality. The tone mapping information SEI message provides information to enable remapping of the color samples of the output decoded pictures for customization to particular display environments. See the paper below.
G.J. Sullivan et al, “ New standardized extensions of MPEG-4 AVC/H.264 for professional quality video applications”, IEEE ICIP, I, pp. 13-16, 2007.

(8)Video Annotation in H.264 (3/21/07)

In the Marrakech meeting, JVT decided to create the AHG on Video Annotation SEI message. The main goal of Video Annotation AHG (ad hoc group) is to study some potential issues on making the compressed video bit stream have more functionality beyond compression, e.g., to support fast video search, value-added applications, or content management. Two reference proposals are JVT-U059 and JVT-V060.
Here we would like to initiate the discussion on this topic. All who are interested in this topic are welcome to make comments in this email thread or contact the chairs or me. Thanks! 
AHG Mandates:


  • Identify applications for video annotation and their requirements

  • Work out suggestions for support needs in AVC

  • Find/create test material

  • Define experiments

Best regards,

Quqing Chen




TRANSCODER

Design, develop, implement and evaluate

H.264 to VP6 Transcoder (Jay Padia has implemented this as his M.S. thesis. May 2010.

(see C.Holder and H. Kalva, “ H.263 to VP6 transcoder, “ SPIE, vol. 6822 (VCIP), San Jose, CA, Jan. 2008. Access from http://spiedigitallibrary.aip.org) Pl access reference 1 from this paper (web site on flash8)

G.F.-Escribano et al, “ An MPEG-2 to H.264 video transcoder in the baseline profile”, IEEE Trans. CSVT, vol. 20, pp. 763-768, May 2010.




Wyner-Ziv to H.264 transcoder (see E. Peixoto, R.L. de Queiroz and D. Mukherjee, “ Mobile video communications using Wyner-Ziv transcoder”, “ SPIE, vol. 6822 (VCIP), San Jose, CA, Jan. 2008. Access from http://spiedigitallibrary.aip.org) . (several papers on Wyner-Ziv coding in this volume).
Subramanya has worked om WZ-codec as his M.S thesis

IEEE Trans on CSVT, Vol.18, April 2008. See the paper below. Warped DCT can this be extended to video coding?



Top of Form

Parameter Embedding Mode and Optimal Post-Process Filtering for Improved WDCT Image Compression
Urhan, O.; Erturk, S.
Page(s): 528-532
Digital Object Identifier 10.1109/TCSVT.2008.918769
AbstractPlus  | Full Text: PDF (168 KB)
Rights and Permissions




Bottom of Form







  • Enhanced AC-3 standard of ATSC (see the paper cited in ATSC DTV standard)

  • MPEG 1 Layer3 (MP3), MPEG-2 AAC and HE-AAC. Discuss encoder decoder block diagrams, advantages and disadvantages.

  • MPEG Surround audio coder, "ISO/IE C230003-1 Information Technology - MPEG audio technologies-Part 1, MPEG Surround, Feb. 2007.

  • Latest audio codecs





  • Add DTS, DTS-High Definition High Resolution Audio, DTS-High Definition Master Audio coding standards to the list of Audio Coding Standards list on slide 13. These are the latest audio codecs developed by DTS and are using in the Blu-Ray discs.





  • See: S.-T. Hsiang, "A new subband/wavelet framework for AVC/H.264 intraframe coding and performance comparison with motion-JPEG 2000", SPIE/VCIP, vol.6822, pp. 68220P-1 through 12, Jan. 2008. Implement this new intraframe scalable coding.

  • See R.G. de Oliveria and R.L. de Queiroz "Intra prediction versus wavelets and lapped transforms in an H.264/AVC coder", IEEE ICIP 2008, pp.137-140, San Antonio, TX, 2008. Implement lapped transforms/wavelets in H.264/AVC intra coding and compare the performance.

  • See D. Marpe et al, "Performance evaluation of motion- JPEG 2000 in comparison with H.264/AVC operated in pure intra coding mode", proc. SPIE, vol.5266, pp.129-137, Feb.2004. Implement this performance comparison.

Rahul_panchal@yahoo. Com 6/14/09

As such not much is published about H265. ( It is officially called high efficiency video coding.) All the info about tools which improves H264 and can become potential tool in H265 can be obtained from VCEG/MPEG ftp site.
I do not have document numbers handy, but I can give the list of interesting tools.
(1) Rate Distortion Optimized Quantization (RDOQ)
     This is still H.264 decoder compatible...
(2) Mode Dependent Directional Transform (MDDT)
(3) Bigger blocks (64x64) and Bigger Transform( 16x16,16x8 and 8x16)
 These bigger transform as Integer approximation of DCT designed in Matlab...
 A good thesis topic can be "A better design of bigger transform for H265".
(4) Adaptive Interpolation Filter (AIF)
 Drawback: It needs at least 2 encoding pass to design the filters. So its more computationally complex.
 Currently these are 6 different flavors of AIF in the KTA (key technical areas) software.
 The latest proposal was about single-pass AIF, but there they designed for P frames which is easy to digest.
 As of now, no idea how to design for B frames using single pass.
 Potential PhD thesis "Design of AIF using single pass encoding".
(5) Competing to AIF is HPF (High Precision Filter) which is computationally simpler than AIF.
(6) Quadtree Adaptive Loop Filter (QALF)
 Competing to QALF is "Post Filter" which unlike QALF is out of the encoding loop.
 Thesis Topic: "Simplification of QALF without losing much of coding gain"
(7) Internal Bit Depth Increase (IBDI)
(8) Motion Vector Competition
(9) New Bi-Predictive Intra modes.
(10) 1/8 pel motion vector ( I do not see this topic as interesting 265 latest developments)
G. Sullivan, “ Recent developments toward standardization of next generation video coding”, SPIE Optics+Photonics, vol. 7798 (paper 30), SanDiego, CA, Aug. 2010. This can lead to several thesis topics/projects.

H.264 open source SVC decoder library 6/18/09 SCALABLE VIDEO CODING


I would like to announce an open source SVC decoder library we have developed that has been later on integrated in 2 different open source players (TCPMP and Mplayer).


This decoder was initially designed inside a French national project called scalimages.
The source code of this decoding library is available here: https://sourceforge.net/projects/opensvcdecoder/
You can find further information at https://sourceforge.net/apps/mediawiki/opensvcdecoder/index.php?title=Main_Page on the installation, features and additional tools related to our SVC decoder library.
The library performance is up to 50 times faster than the JSVM and supports up 2 enhancement layers (however we can change it easily)
The SVC decoder is conformant to the following sequences http://opensvcdecoder.sourceforge.net/JVT-AB023.xls  (see IETR conformance entry in this tabular).
This decoder has been also ported over several platforms such as PDAs and DSP from TI. It will serve as a basis for future development in MPEG RVC (Reconfigurable Video Coding).
If some people are wondering to contribute to this library, they are welcome.

Best regards,

______________________________________________
Mickaël RAULET
Ingénieur de Recherche/Research Engineer

Institut d' Electronique et de Télécommunications de Rennes (IETR) UMR CNRS 6164

Tél : +33 (2) 23 23 82 83
Fax : +33 (2) 23 23 82 62
Port : +33(6) 81 08 35 66

IETR/Groupe Image


INSA de Rennes
20, avenue des Buttes de Coësmes
35043 Rennes Cedex

Dear jvt-experts, 6/19/2009


A new win32 release and source code of mplayer integrated our open svc decoder has been released (version 1.1 http://sourceforge.net/projects/opensvcdecoder/ ). This version integrates layer switching on mplayer and will keep the higher available resolution when displayed. For the layer with finer resolution it will be upsampled. You can also find test sequences to download here.

http://sourceforge.net/project/showfiles.php?group_id=263634&package_id=324062
Mplayer can be used as follow to switch from one layer to another (hotkeys b (to go up), c (to go down)).

http://sourceforge.net/apps/mediawiki/opensvcdecoder/index.php?title=Players_Overview#Mplayer
Another suitable alternative is to use the following command line on the raw stream to select the desired layer (following this DQ_Id = dependency_id<<4+quality_id).

mplayer –fps 25 –layer 16  name_of_the_raw_stream (where DQ_Id is 16 here).


As mentioned before, the library has been designed using a dataflow graph tools and has been ported over several platforms. Our first goal was to implement the code onto a DSP from TI (you can find some TI optimized functions in the developed version) or a processor ARM and Xscale of a pocketpc. Finally our reference code was finally good enough on a PC implementation for the national project purpose.
This dataflow description will serve as a basis of the MPEG RVC Framework I am taking care of (http://www.chiariglione.org/mpeg/visions/rvc/index.htm ). The MPEG RVC framework is based on a dataflow programming Language called CAL (www.ece.umd.edu/DSPCAD/papers/bhat2008x4.pdf, http://opendf.wiki.sourceforge.net/) whose goal is to reuse commonalities between MPEG decoders and to help the description of future standard.
This open svc decoder description will be the first analysis of a CAL dataflow program that we are currently and that we expect to be easily scalable to multicore.
The decoder has been tested with success on the conformance testing of the JVT, with the version 9.16 (sequences are here [http://wftp3.itu.int/av-arch/jvt-site/bitstream_exchange/SVC/]).
For devel purpose, you can follow the wiki [http://sourceforge.net/apps/mediawiki/opensvcdecoder/index.php? title=Main_Page] for installation, features and devel part of the SVC decoder.
As mentioned before it has been built into two players, TCPMP and mplayer, just to make demos. To be able to decode SVC stream within TCPMP, we have modified mp4box to make it compliant with the SVC amendment. This modified version is downloadable from our sourceforge project. Using this modified version is strongly advised to encapsulate SVC stream, in order to get a compliant with SVC streams.

Best regards,


Mickaël Raulet

JBIG JOINT BINARY IMAGE GROUP



http://www.cl.cam.ac.uk/~mgk25/jbigkit/
implementation of JBIG standard
No reference image quality assessment
See paper by R. Hassen, Z. Wang and M. Salama, “ No reference image sharpness assessment based on local phase coherence measurement”, IEEE ICASSP2010, pp.2434-2437, March 2010. Implement this and improve on this further.
SSIM in RGB Color Spaces

http://books.google.com/books?id=1PGC8ju5ytcC&pg=PT73&dq=metric+better+than+SSIM#v=onepage&q=metric%20better%20than%20SSIM&f=false

SSIM in complex Wavelet Domain

http://www.cns.nyu.edu/~zwang/files/papers/icassp05.pdf

INTEGER TRANSFORMS

See the paper J. Dong et al, “2-D order integer transforms for HD video coding”, IEEE Trans. CSVT, vol. 19, pp. 1462-1474, Oct. 2009 and the references cited.

Several large size INTDCTs are proposed in HEVC (see files INTDCT3, INTDCT4,INTDCT5). Compare their performances, transform coding gain. Variance distribution, normalized basis restriction error, normalized mse etc.

Several research topics, specially implementing/evaluating AVS video enhanced profile (EP) and H.264 high profile (HP) based on 2-D order 16 integer transforms (both NICT and MICT) are suggested. See Section V – Integration of the 2-D order-16 integer transforms into the standards – of this paper.

RELATED PAPERS ON LARGE SIZE TRANSFORMS

R. Joshi, Y.A. Reznik and M. Karczewicz, “ Efficient large size transforms for high-performance video coding”, SPE Optics + Photonics, vol. 7798, San Diego, CA, 1-3, Aug. 2010.

A.T. Hinds, “ Design of high-performance fixed-point transforms using the common factor method”, SPE Optics + Photonics, vol. 7798, San Diego, CA, 1-3, Aug. 2010.

W.K. Cham, “Simple order-16 integer transform for video coding” IEEE ICIP 2010, Hong Kong, Sept.2010.

Suvinda (EE5359 Fall 2010 projects) has extended this to 32x32 INTDCT.

Implement these large size transforms in the proposed HEVC (high efficiency video coding)

FINGERPRINT ENHANCEMENT

In [LA1] Raičević and Popović (please access the paper) apply adaptive directional filtering in the frequency domain for fingerprint image enhancement and as well denoising. Block diagram for the fingerprint enhancement is shown in Fig. 4. Note that this is an application of FFT. Figures 6 and 7 show directional filtered images and enhanced image respectively. Apply this algorithm to smudges and corrupted fingerprints (see Fig. 1) and obtain the enhanced images (similar to Fig. 7). Filtered images for directions (a) 22.5 degrees and (b) 90 degrees are shown in Fig. 6. Obtain filtered images for other directions besides 22.5 degrees and 90 degrees such as 45 degrees, 67.5 degrees, 112.5 degrees, 135 degrees etc. Write a detailed project report based on your simulations. Review of the references listed at the end of this paper will be very helpful.


  1. A.M. Raičević and B.M. Popović, “An effective and robust fingerprint enhancement by adaptive filtering in frequency domain,” Series: Electronics and Energetics, Facta Universitatis (NIS), Serbia, pp. 91–104, April 2009.

Available: http://factaee.elfak.ni.ac.rs/

ramesh.rao@comcast.netcoders. See M. Bosi and R.E. Goldberg, “ Introduction to digital audio coding standards”, Norwell, MA: Kluwer, 2002. (reviewed in EURASIP Newsletter, vol.15, pp.7-8, March 2004.) A. Spanias, T. Printer and V. Atti, “ Audio Signal Processing and Coding “, Hoboken, NJ: Wiley, 2007.

High performance video coding   HVC   (OCT. 2009)
CFP (call for proposals) To develop new standards beyond H.264/MPEG-4 Part 10
I have this file. ISO/IEC (MPEG) and ITU-T (VCEG) have combined together with the goal to develop new video coding standards that will have superior performance compared to H.264/MPEG-4 Part 10.

Universal Image Quality Index (UIQI)


  1. Z. Wang and A.C. Bovik, “ Universal Image Quality Index” IEEE SP letters, vol. 9, pp. 81-84, March 2002

  2. Y-G. Wang et al, “Robust dual watermarking algorithm for AVS video”, SP:IC, vol. 24, issue 4, pp. 333-344, April 2009. Can be accessed from SEL using Science Direct. (SP:IC Signal processing: Image communication)

Read the first paper in detail especially equations 1-3. See figures 1 and 2. Using Lena and other images such as baboon, mobile calendar, foreman and, bus, obtain various corrupted/contaminated images. List Q for the various distortions (see table 1).
Evaluate SSIM for all these corrupted images. Prepare a table similar to Table 1 (paper # 1) showing type of distortion, MSE, Q and SSIM. What are your conclusions?

Hint: Access the MATLAB software given in the link below:

http://www.cns.nyu.edu/~zwang/files/research/quality_index/demo.html
DUAL WATERMARKING

Y-G. Wang et al, “Robust dual watermarking algorithm for AVS video”, SP:IC, vol. 24, issue 4, pp. 333-344, April 2009. Special issue on AVS and its applications pp. 254-344. Can be accessed from SEL using Science Direct. (SP:IC Signal processing: Image communication) see all the references at the end of this paper.

Embedding one watermark in the luminance and another in the chrominance (both in the compressed domain) based on the JND concept of the HVS have led to robustness against all types of attacks. Also the watermark detection process has low computational complexity. JND: just noticeable distortion, HVS: human visual sensitivity.

Implement the robust dual watermarking algorithm proposed in this paper. Use different watermarking images (see Fig. 9).

Can this technique be extended to H. 264, DIRAC and VC-1? i.e., robust dual watermarking algorithm technique? These are good research (thesis) topics.

MULTIPLE DESCRIPTION CODING




  1. Radulovic et al, “Multiple description video coding with H.264/AVC redundant pictures”, IEEE Trans. CSVT, vol. 20, pp. 144-148, Jan.2010.

See also references cited at the end of this paper.
2. V. K Goyal, “ Multiple description coding: Compression meets the network”, IEEE SP Magazine, vol. 18, pp. 74-93, Sept. 2001.
These papers can lead to several projects.

5/31/2010

Moscow State University Graphics& Multimedia Laboratory has finished 6-th H.264 codecs comparison. It is intended for practical researchers and developers in the field of high-end video compression. You can find report here http://compression.ru/video/codec_comparison/h264_2010
We have tested newest implementations of MPEG-4 AVC/H.264 video codecs and compare with XviD (MPEG-4 ASP) and Theora encoders.
7/5/2010 VP8 vs H.264
VP8 has recently attracted a lot of interest after it was owned by

Google. As you know, on May 2010, the WebM Project was launched,

featuring contributions from "Mozilla, Opera, Google and more than

forty other publishers, software and hardware vendors" in a major

effort to use VP8 as the codec for HTML5.
As one of appendixes to the annual H.264 comparison report

an additional VP8 encoder vs. x264 encoder comparison was presented.


We have tested VP8 encoder and compare its encoding quality and

speed with x264.


The final report contains all RD-curves, bitrate handling analysis and

speed/quality graphs. Six different VP8 presets were tested which

were chosen with the help of VP8 developers (so those were VP8

developers guided settings).


http://www.compression.ru/video/codec_comparison/h264_2010/vp8_vs_h264.html
Please let us know, if any questions using

videocodec-testing@graphics.cs.msu.ru
Yours,

Dr. Vatolin

Adaptation of vision model to JPEG 2000 image coder (slide 43 in DVIQPCPPT).
Embed the vision model in JPEG2000 and compare its performance with the original

JPEG2000 in terms of mse, psnr, ssim, uiqi etc, based on standard test images.


VP9/VP10 Adobe Flash (0n2 technologies)


http://blog.webmproject.org/2010/06/future-of-vp8-bitstream.html
http://www.webmproject.org/code/repository-layout/
http://review.webmproject.org/#change,56
The closest thing there is to VP9/VP10 at this point is the WebM experimental branch which is open sourced and freely available here (I have included some other helpful links as well):

Timothy Reusing [treusing@google.com] 2/4/2011


HARDWARE ORIENTED FAST MOTION ESTIMATION (review the paper below and the references)


O. Ndili and T. Ogunfunmi, “ Algorithm and architecture co-design of hardware-oriented modified diamond search for fast motion estimation in H.264/AVC”, IEEE Trans. CSVT, accepted. (Feb. 2011).
Yüklə 193,76 Kb.

Dostları ilə paylaş:
1   2   3




Verilənlər bazası müəlliflik hüququ ilə müdafiə olunur ©muhaz.org 2024
rəhbərliyinə müraciət

gir | qeydiyyatdan keç
    Ana səhifə


yükləyin