Liaison report on imagerelated activities in SC37 and activities in JTC1 SC29 WG1 and WG11

1
Liaison report on image-related activities
in SC37 and activities in JTC1 SC29 WG1 and
WG11
2
Resolutions from JTC1 SC37 WG3 meeting

• Israel, January 2008

3
WG3 resolutions relevant to JPEG

• RECOMMENDATION 3.11 Call for Contributions on
  19794-4 Revision 1
• SC 37 WG 3 requests its Secretariat to issue a
  call for National Body contributions that
  investigate the effect of varying compression
  levels on quality of fingerprint images or
  matching performance in ISO/IEC 19794-4.
• Contributions received by 22 May 2008 will be
  considered at the July 2008 WG 03 meeting

4
WG3 resolutions relevant to JPEG/SC24

• RECOMMENDATION 3.12 Call for Contributions on
  19794-5 Amdt. 2
• SC 37 WG 3 requests its Secretariat to issue a
  call for National Body contributions that further
  clarify the use of X3d, especially regarding the
  following topics
• a) size of the representation
• b) support of grayscale images
• c) how to integrate the error map
• d) validity of the point map representation
• e) could the range map also be covered by using
  X3d
• NOTE The colour mapping is incorrect in the
  current contribution (attached to SC 37 N 2376)
• in ISO/IEC 19794-5 PDAM 2 Biometric data
  interchange format - Part 5 Face Image Data AM
  2 Three Dimensional Face Image Data Interchange
  Format (see SC37 N2376 DE 19)
• Contributions received by 22 May 2008 will be
  considered at the July 2008 WG 03 meeting

5
WG3 resolutions relevant to JPEG

• RECOMMENDATION 3.13 Call for Contributions on
  19794-5 Revision 1
• SC 37 WG 3 requests its Secretariat to issue a
  call for National Body contributions on the
  following topics in ISO/IEC 19794-5 (Revision)
  Biometric data interchange format - Part 5 Face
  image data.
• a) inclusion of video data (e.g., short sequences
  of video frames)
• b) prohibiting or discouraging image-altering
  techniques (e.g., cropping, contrast enhancement,
  etc.)
• c) adding a field recording illumination
  characteristics at image-capture time
• d) measuring the 2mm resolution on
  already-prepared photographs, specifically
  regarding scanning
• e) unacceptable facial expressions not already
  included in SC 37 N 2377, AT 18.
• f) handling databases of test / reference data
• g) specification of minimum sizes (in pixels) of
  the facial images, based on interocular distances
• Contributions received by 22 May 2008 will be
  considered at the July 2008 WG 03 meeting in
  Korea.

6
WG3 resolutions relevant to JPEG

• RECOMMENDATION 3.14 Call for Contributions on
  19794-6 Revision 1
• SC 37 WG 3 requests its Secretariat to issue a
  call for National Body contributions on the
  following topics in ISO/IEC 19794-6 (Revision)
  Biometric data interchange format - Part 6 Iris
  image data.
• a) on interoperability tests related to alternate
  compact formats
• b) on results of a study on iris image
  compression. WG3 strongly supports the proposed
  IREX 08 tests by NIST and would encourage NIST to
  examine both published compact formats together
  with proposed new formats, such as ROI
  compression.
• c) on allowable intra-iris Signal/Noise ratio
  value
• d) on effect of using available ROI
  implementations in COTS JPEG 2000 compressors
• e) on differences in efficiency between JPEG and
  JPEG 2000 at various compression ratios
• f) on a study incorporating the measurement
  method as described in US-12 of N2386rev
• g) on cautioning text for the use of optional
  parameters in the image header about pupil
  centre, diameter of the pupil, iris centre and
  diameter of the iris.
• h) on compact formats
• Contributions received by 22 May 2008 will be
  considered at the July 2008 WG 03 meeting in Korea

7
Update on JPEG activity(JTC1 SC29 WG1)

• San Francisco meeting
• 70 delegates from 13 national bodies
• March 2008

8
Progression of JPEG-XR

• Reference software
• New C code developed by Picture Elements see
  N4560 with explanatory slides in N4567
• File Format
• Agreed to unbundle from codestream
• Reflected in new patent statement from MS as
  N4582
• Will define profile of ISO media file format that
  can carry JPEG XR
• Camera Industry
• Presentation from Megachips (chip manufacturer)
  who have implemented JPEG, JPEG XR and JPEG 2000
  and compared their performance N4537
• Progress at next meeting (new V2 CD as draft text
  for FCD to issue in Poitiers)

9
Resolution 43 Restructuring of JPEG Digital
imaging systems and JPEG XR

• WG 1 is recommending the restructuring of JPEG DI
  and JPEG XR into two work items. Under this
  recommendation JPEG XR will continue under the
  current work number ISO 29199 (changing the name
  to JPEG XR) and a new work item for JPEG DI will
  be introduced. WG 1 has identified possible
  sub-divisions of JPEG DI (currently the System
  Overview is in WD stage for JPEG DI Part 1) and
  is requesting comments on these recommendations
  and possible new work items. The following are
  possible sub-divisions of the JPEG DI standard
• Systems overview (Technical Report TR)
• JPEG family of compression standards (TR)
• File Format/Metadata (TR)
• JPEG DI Interactive Protocols (client/server)
  (IS)
• JPEG DI Security (IS)
• Evaluation methodologies for assessing
  compression technology (TR)
• JPEG DI Application profiles (TR)
• WG 1 requests National Bodies to comment on the
  proposed restructuring of the JPEG DI and JPEG XR
  standards. Specifically, WG 1 requests each
  National Body to comment on the recommended work
  items, new work items that should be considered
  in JPEG DI, National Body support for the new
  work items, and recommendations for working

10
Other items of note

• Request for Category C Liaison between WG 1 and
  the International Color Consortium N4570
• Presentation on patent directives N4593
• Presentation on JPEG Historical Archive N4547
• Study text for new work item using JPEG2000 in
  Broadcast Applications N4580
• JPEG Broadcast Backhaul Applications N4581
• Working Draft (WD) to 15444-1 Amendment 3
  guidelines for digital cinema applications N4566
• Text of ISO/IEC WD 24800-4 (JPSearch File Format)
  N4576
• Working Draft of JPSearch part 5 - Data
  Interchange Format between Image Repositories
  N4590
• JPEG XR / JPEG 2000 Coding Complexity profiling
  and PSNR analysis N4586
• Proposal for objective distortion metrics for AIC
  standardization N4548

11
Resolution 33 Approval of JPSearch Workplan

• WG 1 approves and re-affirms the following work
  plan for ISO/IEC 24800 JPSearch
• Part Title WD PDTR DTR TR
• 1 System framework and components 05/07 06/07
  07/04 07/07
• Part Title WD CD FCD FDIS
• 2 Schema and ontology registration
• and identification
  08/07 08/10 09/04 09/07
• 3 JPSearch Query Format 08/07 08/10 09/04
  09/07
• 4 Metadata embedded in image data
• (JPEG and JPEG 2000) file forma)t 08/03
  08/07 08/10 09/04
• 5 Data interchange format between
• image repositories 08/03 08/10 09/04 09/07
  

12
Resolution 56 WG 1 affirmation of IP guideline
statement submission

• WG 1 re-affirms its support for its IP guideline
  statement first approved at the 16th WG 1 Seoul
  Meeting, March 1999.
• WG 1 requires all participants within all
  National Bodies to disclose and identify any and
  all patent rights and the specific technologies
  within the Verification Model to which they
  apply. Further, WG 1 requires this disclosure and
  identification at the time of submission of
  technology for VM consideration if submitted by
  the patent holder or no later than one meeting
  after submission of technology if the technology
  is not submitted by the patent holder. Further,
  WG 1 requires that the form contained in WG 1
  N1267 be completed as part of this disclosure.
  This request is in accordance with Common Patent
  Policy for ITU-T/ITU-R/ISO/IEC, clause 3.
• The relevant ISO/IEC directives are contained in
  the integration group report, WG1N1262.

13
Resolution 57 WG 1 affirmation of SC 29
resolution on IPR guidelines - 1

• WG 1 complies with the Resolution 9 SC 29 IPR
  Guidelines of the 20th SC 29 Plenary Meeting,
  2007-04-30/05-01, San Jose, USA
• SC 29 affirms and supports Common Patent Policy
  for ITU-T/ITU-R/ISO/IEC that requires a party
  that submits a technology contribution for
  proposed inclusion in any SC 29 standard to
  disclose to the WG the existence of patents or
  published pending patent applications, hereafter
  referred to as "IP rights", of which the party is
  personally aware and which the party considers to
  cover any item of its contribution.
• "Clause 3, Guidelines for Implementation of the
  Common Patent Policy for ITU-T/ITU-R/ISO/IEC any
  party participating in the work of the
  Organizations should, from the outset, draw their
  attention to any known patent or to any known
  pending patent application, either their own or
  of other organizations."

14
Resolution 57 WG 1 affirmation of SC 29
resolution on IPR guidelines - 2

• SC 29 affirms the Common Patent Policy for
  ITU-T/ITU-R/ISO/IEC of only considering for
  inclusion in a standard (unless the ISO Council
  authorizes an exception) technology that is
  either 1) free of IP rights, or 2) available to
  all implementers of the standard without a
  royalty or license fee and under other reasonable
  and non-discriminatory ("RAND") terms and
  conditions, or 3) available to all implementers
  of the standard for a reasonable royalty/license
  fee and under other RAND terms and conditions.
• Although royalty-bearing patented technologies
  may be included in SC 29 standards, SC 29
  suggests to its WGs to promote, whenever
  possible, the inclusion of technologies that
  either do not require a patent license, or that
  only require a RAND license without a royalty or
  license fee. Additionally, SC 29 encourages
  disclosure by the IP rights holder of the type of
  licensing under which these IP rights will be
  made available for conforming implementations of
  ISO/IEC standards after final publication -- that
  is either 1) royalty/license fee-free and under
  other RAND terms and conditions, or 2) reasonable
  royalty/license fee and under other RAND terms
  and conditions. Negotiations regarding specific
  licensing terms and conditions are left to the
  parties concerned and are performed outside of SC
  29. Further, SC 29 requires software and
  technology contained in any party's contribution
  to an SC 29 standard to be made available by such
  party to all members of the SC 29 WG developing
  such standard, on a royalty- and license fee-free
  basis, solely .

15
Future WG1 meetings

• 45th 2008 Jul 7 - 11 Poitiers, France
• 46th 2008 Oct 13 17 Pusan, Korea
• (co-located with WG11)
• 47th 2009 Apr 20 24 Maui, Hawaii
• (co-located with WG11)
• 48th 2009 Jul Sardinia, Italy (to be
  confirmed)
• 49th 2009 Oct 26-30 Xian, China
• (co-located with WG11, TBD)
• 50th 2010 Apr Perth or Cairns, Australia
• (co-located with WG11,

16
Update on MPEG activity(JTC1 SC29 WG11)

• Antalya meeting January 2008
• Next meeting Archamps
• 28 April 2nd May

17
Image Signature

• The ImageSignature descriptor provides a
  "fingerprint" to uniquely identify individual
  visual media items. It has no direct association
  with specific visual features. The ImageSignature
  descriptor (the signature) is robust (unchanging)
  across a wide range of common editing operations,
  but is sufficiently different for every item of
  "original" content to identify it uniquely and
  reliably just like human fingerprints.

18
Image Signature

• The descriptor is able to distinguish between two
  images which have similar content and two images
  which are the same
• The descriptor analyses lines and circles in an
  image to extract a 512 bit signature that is
  robust to common processing operations that may
  be performed on the image.

19
Trace transform

• The Trace transform applies functionals over all
  of the lines in the image. A functional is a
  real-valued function on a vector space V, usually
  of functions. A line is parameterised by the
  distance d and the angle ? and points on the line
  by the parameter t. The result of applying a
  Trace transform is a 2D function. A further
  functional can then be applied to the columns of
  the Trace transform to give a vector of real
  numbers. This second functional is known as the
  diametrical functional and the resulting vector
  is known as the circus function. The properties
  of the circus function can be controlled by
  appropriate choices of the two different
  functionals (trace and diametrical).
• The trace-transform based technology was promoted
  from Working Draft to PDAM. (ISO/IEC
  15938-32002/PDAM 3)

20
VCE-6 Image Signature (Visual identifier)

• Experiment Conditions
• The images are selected from the CD-ROMs "Art
  Explosion 800000" by Nova, available at
• http//www.amazon.co.uk/Nova-ARW-Art-Explosion-800
  000/dp/B0001XWNSS/refpd_bbs_sr_1/203-3503298-4948
  756?ieUTF8ssoftwareqid1183552443sr8-1
• The experiment consists of two tests, one for
  independence and for robustness

21
VCE-6 Image Signature (Visual identifier)

• Independence Test
• A large database (N135k images in List A) of
  different (independent) images shall be used to
  determine the parameters (e.g. a threshold) for
  each algorithm corresponding to the required
  maximum false positive rates. The maximum rate
  of false positive matches shall be
• For basic modifications only 0.05 ppm (parts
  per million)
• For all modifications 10.0 ppm (parts per
  million)
• All possible pairs of images in the database
  shall be used for this purpose (excluding
  self-comparison) the number of comparisons is
  therefore (Nx(N-1)/2).

22
VCE-6 Image Signature (Visual identifier)

• Robustness Test
• Each algorithm is tested for robustness to
  various types of modifications and its success
  ratio is determined in each category (see Table
  1). For this test a second database of M (10k)
  original images is created according to the
  List B. For each original image from the
  B-list, 41 modified (transformed) versions are
  created, using the transformation parameters from
  Table 1. Software to generate the transformed
  images is available for download from
  ftp//image.dgu.ac.kr

23
VCE-6 Image Signature (Visual identifier)

• The basic modifications are
• Brightness change
• Color to monochrome conversion
• JPEG compression with varying quality factors
• Color reduction
• Gaussian Noise
• Histogram equalization
• Image enhancement via Auto-level
• Blur
• Scaling
• Rotation by angle 90, 180 and 270.
• Flip
• The additional complex modifications are
• Rotation
• Translation
• Aspect ratio change
• Crop
• Skew
• Perspective

24
VCE-7 Video Signature

• It was reported that existing MPEG-7 visual
  descriptors give limited performance in the role
  of video signatures especially when text is
  overlaid on video. This CE explores what
  technologies, if any, are suitable for the task
  of identification of video-clips (i.e. video-clip
  signatures). The main focus is on accuracy and
  robustness of identification. (Note one of
  application of the video identifiers is to detect
  replicas of a video material).

25
RVC Reconfigurable Video Coding

• MPEG-B related core experiment (8 contributions)
• Automatic generation of parser VLD decoding,
  Btype generation
• Compressed bitstream syntax (CDDL)
• MPEG-C related contributions (4 contributions)
• AVC interlaced coding FUs
• VLD decoding FUs
• T oken classification
• MPEG-2 FUs
• There was also one exploration contribution on an
  AVS parser FU, and one contribution on RVC
  support tools status of CAL2C generation tool.
• CDs for both parts (23001-4 and 23002-4) will be
  finally edited and released at the end of the
  week. These will be as approved at the last
  meeting, not including compact binary
  description. FCD is expected in April.

26
Free-viewpoint TV

• The focus of the next phase of the FTV work will
  be 3D TV supporting N-view displays (various
  types) where only a low number of views (say up
  to three) and depth map is encoded. Test
  sequences will be designed accordingly, with a
  parallel camera set-up, and with rectification
  most probably being necessary. Test sequences
  would be provided in rectified, illumination and
  colour compensated form. The maximum angle
  between left and right will be around 20 degrees
  (todays displays 10 degrees).
• A new updated call for test sequences and depth
  estimation software will be issued, including
  high level description of the scope of FTV.
• An investigation into depth estimation and
  interpolation performance will be performed in an
  exploration experiment, based on the algorithm
  proposals submitted. This should bring evidence
  about
• Whether sufficiently food depth estimation
  algorithms are available
• What level of quality can be achieved in view
  synthesis, and whether PSNR comparison against
  original views is useful
• Determine the suitability of the test sequences
  already available and subsequently provided, for
  the purposes of the up-coming call for proposals.

27
JVT Update

• The new edition of 14496-10200x, including the
  Professional profiles and SVC is due to be
  completed in October.
• The conformance and software amendments for the
  Professional Profiles and SVC are in progress,
  reaching FPDAM at 83rd meeting.
• A corrigendum is underway, including corrections
  to SVC.
• The other current work taking place is on the
  following topics
• Multi-view video coding (MVC 14496-10200X/Amd.1)
  
• Study of benefit for possible extension to SVC to
  support bit depth scalability
• There are other proposals relating to SEI
  messages to support slicing and enhancements to
  coding efficiency

28
SVC verification tests

• N9577 Report on SVC Verification tests
• Wide variety of test scenarios for all three
  profiles
• Including teleconferencing, internet broadcast,
  HD broadcast, production applications
• The results indicate that with at most a penalty
  of 10 of bit rate, scalability functionality can
  be supported
• There is a very clear advantage over simulcast
• Conclude that the challenge of scalable video
  coding has finally been met
• Evidence that this standard is useful in the
  market and may quickly be taken up

29
Video Surveillance (VS-AF)

• ISO/IEC 23000 (also known as MPEG-A) is an MPEG
  standard that supports a fast track to
  standardization by selecting readily tested and
  verified tools taken from the MPEG body of
  standards and combining them to form an AF
  (Application Format). If a needed piece of
  technology is not provided within MPEG, then
  additional technologies originating from other
  organizations can be included by reference in
  order to facilitate the envisioned AF.
• The Video Surveillance AF is a file format
  designed to provide for a first level of
  interoperability for video-based surveillance
  systems. The file format provides the overall
  structure for storing video content and
  associated meta-data in a single file It contains
  MPEG-4 AVC baseline profile video and associated
  MPEG-7 meta-data. Usage of other coded video
  formats will be assisted.
• Liaison sent to IEC TC 9 WG43

30
Joint Meeting on IPTV between ITU-T SG16/Q13 and
ISO SC29/WG11 (MPEG)

• Sunday April 27, 2008 at ITU-T
• Agenda
• SG16 Presentations
• MPEG Presentations
• Overview of MPEG technologies Leonardo
  Chiariglione
• Multimedia Middleware (M3W) Jean Gelissen
• Digital Items (DIs) Xin Wang
• DAC profile of Rights Expression Language (REL)
  Taehyun Kim
• Media Streaming Application Format (MSAF)
  Tiejun Huang
• Digital Media Broadcast Application Format
  (DMBAF) Young Kwon Lim
• Binary Format for Scenes (BIFS) and Lightweight
  application scene representation (LASeR) Young
  Kwon Lim/Olivier Avaro
• Web, IP and Mobile (WIM) TV Olivier Avaro
• Discussion on future collaborations