Horst Tappert - That Rest Prop

Tart Prophets "Fatals Picards" have foreseen this day some time ago with their silly song "A l'enterrement de Derrick". Stephan Derrick, alias Horst Tappert, died on saturday, December 13th.

Fatals Picards - L'enterrement de Derrick

Smile in the sky - My God, it's full of stars

Funny sign in the sky in this cold and damp December: a crescent Moon and two "stars" have just offered us a cosmic smiley. The two brilliant "stars" are our sister planets Venus and Jupiter. This smiling sky has especially being reported from the chinese side of the world, under different angles:

China economy Net
Chine informations
Lhasa: The sky smiles in the night

and last a Youtube video. Maybe a smiling sign from 2001: A Space Odyssey Star-Child, for his 20th birthday.

Major en *let - In honor of Jean Morlet

Jean Morlet (or the Major en -let, in french in the text), one of the fathers of all wavelets, along with some precursors, as in the nice notice Precursors in mathematics: early wavelet bases, by Hans G. Feichtinger, is honored at the end of October in Marseille, France.

A preliminary program below the Morelet crocodile:
October 27 - 29th, 2008
Continuous wavelet transform and Morlet's wavelet, 1978-2008 : International colloquium in honor of Jean Morlet
Organizer: G. Saracco (CNRS, CEREGE Aix-en-Provence)

Centre International de Rencontres Mathematiques (CIRM), Campus de Luminy, Marseille, France

*Preliminary Program*:

1-Alex Grossmann: «Continuous wavelet theory and Morlet's wavelet» (Hommage à J. Morlet) , (CNRS-Evry)

2-Y. Meyer, membre de l'Académie des Sciences: «Des ondelettes continues en géophysique
aux bases orthogonales d'ondelettes» (hommage a J. Morlet) (ENS-Cahan)

3- Pierre Goupillaud : PR, Two superb mathematicians who's innovations were the product of asking question that others had overlooked. (Hommage à J. Morlet), (SEG., USA)

5-Patrick Flandrin, DR: Wavelets, surrogates and non-stationarities,(ENS-Lyon)

6-Stephane Jaffard, PR: Directional continuous wavelet transforms and application to directional smoothness of functions (Paris XII)

7-Ronald Coifman, PR: Wavelet and diffusion geometries on seismic data sets (Yale Univ., USA)

8-Alain Arneodo, DR: Surfing on the genome: a tribute to J. Morlet (ENS-Lyon)

9-Jean-Pierre Antoine: The wavelet transform on the sphere: continuous vs. discrete (Louvain, Belgique)

10-Jean-Luc Stark, : Compressed sensing in astronomy (CEA-Paris)

11-Nick Kinsbury: Complex-valued wavelets, the dual tree and the Hilbert Transform: Why these lead to
approximate shift invariance? (Cambridge Univ., GB)

12-Thierry Paul, DR: Non linear continuous wavelets and non linear coherent states (ENS-Cachan)

13-Ginette Saracco: Multi-scale tomography of buried magnetic or electrical sources: Its localization and characterization.
Application to archaeological structures or volcanic system (CNRS-CEREGE, Aix en provence)

14-Gregory Beylkin, PR: On approximation by Gaussian and its applications (Boulder Univ. Co, USA)

15-Marie Farge, DR: Continuous wavelet analysis of vortex bursting in turbulent flows (ENS-Paris)

16-Matthias Holschneider, PR, Directionnal Poisson wavelets on the sphere (Potsdam, Allemagne)

17- Jean-Claude Risset, DR emerite, medaille d'Or du CNRS, Morlet wavelet are good to hear (CNRS-LMA, Marseille)

18- Peter Frick, PR, Double Wavelet analysis: Method for recognizing stellar activity peculiarities (Moscou Russie)

19- Pascal Sailhac, From seismic wavelets to wavelet transform: using Berlage wavelets to the detection
and characterization of damped transient waves occurring in geophysical time-series", (IPG-Strasbourg)

20- Nicolas Thouveny, PR, Extraction of frequency modulation laws from Earth's magnetic field intensity records
by complex continuous wavelet analysis; contribution to understanding the geodynamo behavior. (CEREGE, Univ Aix-Marseille, Aix en Provence)

21- Say Song Goh, PR, Uncertainty Principles for the Continuous Wavelet Transform, (Rep Singapore)

22- Maurizio Fedi, Discrete and Continuous Wavelet Transform of potential fields with different choices of analyzing wavelets, (Naples Univ., Italie)

23- Roddam Narasimha, PR, Morlet Wavelets and the Solar Connection with Indian Monsoon Rainfall, (Bingalore, Inde)

24- Albert Cohen, PR, Adaptive multiresolution analysis based on anisotropic triangulations, (Univ. Pierre & Marie Curie, Paris)

25- Agissilaos Athanassoulis, Strengthening semiclassical approximations with the use of coarse-scale representations, (ENS-Ulm, Paris)

26- Nele de Shepper, Multi-dimensional continuous wavelet transforms and General Fourier transforms in Clifford Analysis, (Ghent Univ, Belgium)

Featured on the SIVA conferences web page.

Concern fees - Upcoming conferences - PCS 2009

The 2009 Picture Coding Symposium has been announced and its website made public
Website : PCS 2009
https://redpill.ecn.purdue.edu/~pcs2009/
Submission deadline: November 21 , 2008
Conference dates: May 6-8, 2009

Topics of interest include:

• Coding of still and moving pictures
• Content-based and object-based coding
• Scalable video coding
• Coding of multiview video and 3D graphics
• Modeling and synthetic coding
• Virtual/augmented reality and telepresence
• Coding for mobile, IP and sensor networks
• High fidelity visual data processing and coding
• Analysis for coding and adaptation
• Transcoding and transmoding
• Joint audio and visual processing and coding
• Subjective and objective quality assessment metrics and methods
• Error robustness, resilience and concealment
• Coding and indexing for database applications
• Protection and integrity of visual data
• Persistent association of information to visual data
• Joint source and channel coding
• Implementation architectures and VLSI
• New applications and techniques for visual data coding
• Standards for visual data coding

SIVA Conferences

If Mike's dog really ate his frog - On the peer review origin

I originally underestimated the peer review argument in Attempts to hedge out peer review, where Eric Weinstein associates it with the "If Mike's dog really ate his frog" growing meme. In most of the scientific world, we revere pi, setting high standards to what should be published and known. Thus, let's unveil the truth: the dog's name was Bagel, which could become a peer review paradigm, as Lenna is for image processing (another Lenna tells us about a very interesting bagel and salad diet, giving us the missing link between two paradigms).
Chad Orzel quotes this anectode in Uncertain principles, with Google slightly giving it more echoes (a thought to Richard Wright by the way), as pointed out on Nuit Blanche by Igor Carron. He also refers to the soft pion theorem, but what about very cold neutrons?

Concern Fees - About PIMRC 2009

Apparently, from the logs of the SIVA Conferences web page (post title anagramatic explanation), tons of radio people are looking for the next 2009 IEEE International Symposium on Personal, Indoor and Mobile Radio Communications Conference (PIMRC the 20th), which was planned in Yokohama (Japan). Luckily, PIMRC 2009 website has just been released, and holds in Tokyo (Japan again). Don't ask me why. Its 2009 theme is "Giga bit Wireless for Real Personal, Indoor and Mobile Radio Communications".

You may look at the 20th PIMRC 2009 call for papers. The submission dead-line is 06/03/2009 (03/06/09 American style, which looks arithmetically nicer but has ambiguities), and the conference is from 16/09/2009 to 18/09/2008. Topics follow:

PHY related topics
Antenna & Propagation
Modulation & Coding
Equalization and Synchronization
MIMO
Cognitive radio
Millimeter wave
WLAN, WPAN, BAN
Enabling technologies
Beamforming antenna, MEMS, RF CMOS, etc
MAC related topics
Ad-hoc network management
Resource management
Cross-layer management
802.11/15/16 MAC
3GPP/3GPP2 MAC
Network related topics
Cognitive wireless network
Cooperative network
Ecological wireless network
P1900/802.22
Multi-hop network
Mobile agent
Wireless ad-hoc network
Sensor network
Local dependent network
Mobile computing
Mobile IP network
Application related topics
Next generation digital home network
Secure network
Wireless robotics
Social wireless networks
Security
ITS
In-car/Intra-car communications
Millimeter Giga-bit applications
Medical ICT applications

Other signal/image processing people concerned by fees should look at the Signal, Image, Video and Applications conferences page. Submissions welcome.

Well if Mike's dog really ate his frog, let us update

The Mike from "if Mike's dog really ate his frog" is Michael D. Grossberg. After a PhD in Mathematics from MIT, after having his dog being quoted by co-author Yael Karshon, he is apparently active in the signal/image processing community, which is... interesting to me, at least. More interesting, let us have a closer look at Mike's place: www1.cs.columbia.edu/~mdog/. Crypting writings again? No, probably a spur of Mike's Dog. But what would happen if Mike's dog really ate his blog?

What if Mike's dog really ate his frog? - On Liar's grief, faked mythologies

If Mike's dog really ate his frog is taken from a talk by Eric Weinstein: Attempts to hedge out peer review (between slides 73 and 79). It is an if and only if condition which remained long undiscovered (even if cited). It anagrams to Liar's grief, faked mythologies, which is not unrelated to its origin (cryptic writings in a PhD thesis for a student to confound his advisor). An other reference may be found at Project Euclid; just enter "ate + frog" in the fulltext search box, you get this very serious paper:

The moment map and line bundles over presymplectic toric manifolds, Yael Karshon and Susan Tolman, J. Differential Geom. Volume 38, Number 3 (1993), 465-484.

Igor Carron at Nuit Blanche suggested to meme this sentence. Let us join the mob.

P.S.: all apologies to Eric Weinstein for a name misspelling. Instead of an insincere wite (blame), to you I write incense, gold and myrrh gifts in shape of anagrams, hoping to build together irenic new ties with forgiveness, since i renew it: my apology!

A real duet - On dual-tree wavelets (and a Matlab toolbox)

True pieces of art elude science (left picture borrowed from Flickr). Anyway, dual-tree wavelets form a real duet in Hilbert transform type harmony. But the true deal comes from some software implementing this possibly complex wavelet decomposition.

Version 1.2 of the M-band dual-tree wavelet Matlab toolbox has been released. It is available from Caroline Chaux webpage or directly from the link:
http://www-syscom.univ-mlv.fr/~chaux/toolbox/TOOLBOX_DTMband1D_v1.2.zip

The toolbox implements several 2-band and M-band wavelets (e.g. Meyer 2, 3 and 4 bands, Haar, Shannon etc.) in the Hilbert transform based dual-tree wavelet framework. Theoretical results, applications and comparisons may be found in the following papers from
http://www-syscom.univ-mlv.fr/~chaux/publications.html

Image Analysis Using a Dual-Tree M-Band Wavelet Transform
IEEE Transactions on Image Processing, Vol. 15, No.8, Aug. 2006, p. 2397-2412

Noise Covariance Properties in Dual-Tree Wavelet Decompositions
IEEE Transactions on Information Theory, Vol. 53, No. 12, Dec. 2007, pp. 4680-4700.

A Nonlinear Stein Based Estimator for Multichannel Image Denoising
IEEE Transactions on Signal Processing, Vol. 56, No. 8, Aug. 2008, pp. 3855-3870.

Ice Soup - About EUSIPCO 2008 (and signal in general)

The "Signal de Sauvabelin" is a beautiful wooden tower offering a great Laussane panorama. This "signal perspective" was celebrated at the very same place in Switzerland place last week at EUSIPCO 2008. Enjoy some other pictures taken from cold Swiss lakes (hence the Ice Soup) at Picasa.

Meanwhile, i may have another Signal Tower close to home, at La Defense (close to Paris), since Jean Nouvel recently won the contest for this renewal minded building.

Temporary exhibition of the contest's finalists at Cité Chaillot, from 25 June to 30 September 2008.

On my window by 2015. But since then, many other signal and image processing conferences will provide us with travel opportunities. The 17th European Signal Processing Conference, EUSIPCO 2009, conference will be held in Edinburgh, Scotland. An good reason to meet the 32th (Scottish) Signal Squadron? Discover the challenge, as they say. All along the watch tower, there must be some way out of here...

Concern Fees - About conferences (in signal and image processing)

Having spotted before the stricking anagrammatic link between "conferences" and "fees concern", through a digression on Juran's famous hammer and nail quote, this post initiates a series of annoucements on some upcoming signal and image processing conferences of interest since, generally, conference deadlines represent strong incentives for scientific outcomes. Due to the myriad of such events, let us consider this series, to the first order, as a self-remainder. A more complete list of signal and image processing conferences is also available.

First annoucement is SSP 2009 (IEEE Workshop on Statistical Signal Processing). It has been annouced in Cardiff, Wales, from 31/08/2009 to 03/09/2009 (present deadline: 27/04/2009).

The theoretical topics include: Adaptive and nonlinear signal processing, Blind detection and separation, Broadband beamforming, Cognitive signal processing, Detection and estimation, Distributed signal processing, Monte Carlo methods, Multi-modal techniques, Optimization in signal processing, System identification and calibration, Time-frequency and time-scale analysis; with applications in Array processing, radar and sonar, Automotive and industrial applications, Bioinformatics and genomic signal processing, Biosignal processing and medical imaging, Communication systems and networks, Information forensics and security, Multi-dimensional signal processing, New methods, directions and applications, Sensor network.

ICA 2009 (International conference on ICA and Signal Separation) is announced in Paraty, Brazil, to be help in Spring 2009.

Independent Component Analysis and Signal Separation is one of the most exciting current areas of research in statistical signal processing and unsupervised machine learning. The area has received attention from several research communities including machine learning, neural networks, statistical signal processing and Bayesian modeling. independent Component Analysis and Signal Separation has applications at the intersection of many science and engineering disciplines concerned with understanding and extracting useful information from data as diverse as neuronal activity and brain images, bioinformatics, communications, the world wide web, audio, video, sensor signals, or time series.

PCS 2009 (Picture Coding Symposium) takes places in Chicago, Illinois, USA (Purdue University) from 06/05/2009 to 08/05/2009. Deadline to be announced.

Topics of interest inculde: Coding of still and moving pictures, Content-based and object-based coding, Scalable video coding, Coding of multiview video and 3D graphics, Modeling and synthetic coding, Virtual/augmented reality and telepresence, Coding for mobile, IP and sensor networks, High fidelity visual data processing and coding, Analysis for coding and adaptation, Transcoding and transmoding, Joint audio and visual processing and coding, Subjective and objective quality assessment metrics and methods, Error robustness, resilience and concealment, Coding and indexing for database applications, Protection and integrity of visual data, Persistent association of information to visual data, Joint source and channel coding, Implementation architectures and VLSI, New applications and techniques for visual data coding, Standards for visual data coding.

Last for today, ISMM 2009 (9th International Symposium on Mathematical Morphology) in Groningen, The Netherlands, deadline on 16/01/2009, to be held from 24/08/2009 to 27/08/2009. Final website unknown at the present time, first announcement here.

Gross topics: Lattice Theory, Geometry and Topology, Signal Processing, Image Processing, Connectivity, Segmentation, Machine Learning, Algorithms and Architectures, with applications.

Try nerd duel - on a fan of Tyler Durden

Let us welcome a brand new signal processing blog. Le petit chercheur illustré, under the auspices of Fight Club. If you wanna fight, Tyler Durden... Try nerd duel. Le petit chercheur deals today with some ancient form of Matching pursuit. Since its english subtitle is Yet Another Signal Processing and Applied Math blog, its acronym contains the word SPAM (for the Signal Processing and Applied Math part), enough to remind us of the Monty Python's Flying Circus.
But surprise, the whole acronym, YA SPAM, belongs to urban language, as seen in the Urban dictionary. Its definition?

An quick precise reaction to an outburst of words from another person

Maybe someone we already knew for his "not in let multiselectivelet", a somewhat pythonesque wavelet. Very well adapted indeed. Welcome Laurent (yet another one). We hope you quickly grow into a grand chercheur illustre.

Conferences (Concern Fees?) - About hammers and nails

There is some concern about scientific conferences in general, and signal/image/video conferences in particular, as money making machines (hence the fees' concern anagram). Not only places where people meet, share, even copy or steal ideas... Remember the standard quote, attributed to Wilson Mizner (1876-1933):

Copy from one, it's plagiarism; copy from two, it's research

or maybe this quote has also been unfaithfully copied, since it often appear as

If you steal from one author, it’s plagiarism; if you steal from many, it’s research

Steal or copy, two or many. Blame the the cut, correct and paste. Since Mizner's most famous plays is entitled The Deep Purple, shouldn't this sentence be refined along the lines of another solid rock band, Megadeth (Captive honour, from Countdown to extinction):

And when you kill a man, you're a murderer (plagiarist)
Kill many, and you're a conqueror (researcher, then)
Kill them all... Ooh... Oh you're a god!

What would be the research equivalent of "god"? Contributions welcome. Anyway, let us mention a few interesting conferences in signal or image processing. The upcoming EUSIPCO 2008, in Lausanne, Switzerland, along with the satellite international workshop LNLA 2008, dealing with "Local and Non-Local Approximation in Image Processing". Still waiting for the program and its novelties from the sparse or compressive sensing world.

And now for something completely different. About 10 years ago, the department head of an engineering school asked me whether people in the industry did use wavelets? My reply was: "do you teach wavelets to future engineers?" The answer was no. Anyway, there exists a conference dealing with wavelet applications in the industry, organized by Fred Truchetet and Olivier Laligant: SPIE Wavelet Applications in Industrial Processing VI (WAIP 2009), along with SPIE Electronic Imaging in San Jose, California, USA, 16-18/01/2009. Deadline for a short abstract: 16/06/2008.

My concern above was not about wavelets or anything else. But industrial business is often time consuming, cost bounded, and people tend to reuse known tools a lot. Even for weakly related problems. The quote here is the nail/hammer famous line. Though once again, several (digital) versions co-exist; one attributed to Abraham Maslow (1908-1970), founder of humanistic psychology:

If you only have a hammer, you tend to see every problem as a nail

or this one, copied from his book "Psychology of Science":

I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail

Apparently, this idea is good enough to have been used by several persons. Maslow's is said to originate from his mentor Michael Polanyi (1891-1976), while Bernard Baruch (1870-1965) is credited with

If all you have is a hammer, everything looks like a nail

The only one i have been aware of before the world wide web was (in French)

Quand on n'a qu'un marteau dans sa boîte à outils, tous les problèmes ont la forme d'un clou

from a guy simply known as Juran. Which could be translated as:

When you only have a hammer in your toolbox, all your problems are nail-shaped

It might be Joseph M. Juran (1904-2008, he just died on February 28th), management consultant, quality guru. All these folks overlap (in time) and exhibit some redundancy (in some orthogonal direction), which are my two main signal processing concerns. Anyway my reply was about teaching concepts that could help future engineers seize new ideas when needed, to avoid pure repetition.

Another conference is GRESTI 2009 Symposium on Signal and Image Processing, to be held in Dijon, France, from 8 to 11 September 2009. Deadline 15 Feb. 2009. The last announcement for today is tentative, the Mathematical Methods for Natural Images and Textures Processing, organized by G. Peyré, J.F. Aujol and J. Fadili, at IHP, Paris, France, to be help in June 2009.

Swanky Bettors - on West bank Story

May a movie bring some peace in the Middle East? A litte singing, a little dancing, a lot of humm(our)us. A nice feature for this 60th anniversary of the Foundation of the state of Israel.
Swanky Bettors may bet on West bank Story (online video here, and IMDB page there) by Ari Sandel. Official selection of the 2005 Sundance film fest, 28 prizes won in festivals. Love inside inter-family confrontation in a FFF (falafel-fast-food, not FFF, not FFF) world, KK (Kosher-King) against HH (Hummus Hut). David and Fatima replay Romeo and Juliet and West side story. Wanna trade chickpeas for a chic peace?

Cochlear Auxin - Congratulations

Caroline Chaux PhD thesis has been awarded "best thesis in signal and image processing" by the EEA Club, on Tuesday 6th May 2008. The EEA Club is a 40 year old association that gathers teachers and researchers in Electrical and Control Engineering, founding member and part of the European Association for Education in Electrical and Information Engineering (EAEEIE). The official ceremony takes place in Saint-Etienne, France, at the 48ème congrès du Club EEA (www.istase.fr/eea2008/) from May 28 to May 30. Caroline Chaux also received the best student paper award at ICASSP 2005.

The PhD thesis is entitled "Analyse en ondelettes M-bandes en arbre dual ; application à la restauration d’images" or "M-band dual tree wavelet analysis with application to image restoration". Dual tree wavelets represent a special breed of wavelet frames composed of the union of two (M-band) wavelet bases in phase quadrature or forming Hilbert pairs. The sine and cosine functions form a traditional example of Hilbert pairs. Hilbert pairs of wavelets enjoy approximate shift invariance and are low-cost redundant transforms for directional image analysis. They are related to the discrete complex wavelet transform. They have been used for instance in compression, texture analysis, denoising, watermarking... Many others applications are yet to come... and why not on compressed sensing?

A Matlab toolbox for 1-D M-band dual-tree wavelet transforms is made available. Related articles for further reading:

• Nonlinear Stein Based Estimator for Multichannel Image Denoising, IEEE Trans. on Signal Processing, 2008 (to appear),
• Noise covariance properties in Dual-Tree Wavelet Decompositions, IEEE Trans. on Information Theory, 2007
• Image Analysis Using a Dual-Tree M-Band Wavelet Transform, IEEE Trans. on Image Processing, 2006

with applications in the following:

• C. Chaux, P. L. Combettes, J.-C. Pesquet, and V. R. Wajs, "A variational formulation for frame-based inverse problems," Inverse Problems, vol. 23, pp. 1495-1518, June 2007.
• P. L. Combettes and J.-C. Pesquet, "A Douglas-Rachford splitting approach to nonsmooth convex variational signal recovery," IEEE Journal of Selected Topics in Signal Processing, vol. 1, no. 4, pp 564-574, December 2007.
• W. Ayadi, A. Benazza-Benyahia, S. Sevestre-Ghalila, "Quantification of microarchitecture bone junctions based on a dual tree M-band wavelet decomposition", Conf. Proc. IEEE Eng Med Biol Soc. pp. 5634-7; Volume: 2007

Flat Noise Set - on Fast Noiselet transform

After first try to provide some code for the noiselet transform matrix, which yields a flat noise set of coefficients from the Haar system (and to obtain nice pictures), a valuable implemetation has been made public. No lie, fastest Matlab code for noiselets is now available from Laurent Jacques (direct link). David Donoho(*) recently recalled that the compressed sensing litterature shared similarities with concepts in other fiels (an oft essential remark). For instance in data stream processing, as for “heavy hitters” or “Iceberg queries”. The Fast Noiselet is not the nastiest floe in this ocean of litterature.

(*) David L. Donoho, Compressed Sensing, IEEE Transactions on Information Theory, April 2006 (pdf copy).

On complex hadamard matrices

Connections or connexions?

François Morellet is a painter, scultor, engraver, who turned abstract by the 1950's. Some of his works carry strange connections with sparsity, wavelets and even compressive sensing. First, his name, akin to both a crocodile and Jean Morlet, who died almost one year ago. Then, he authored a work entitled "4 trames de tirets simples (non quinconce) 0°-45°-90°-135°", which obviously refer to a set of four oriented separable ("non quincunx") frames, very similar to complex dual-tree wavelets. The remaining terms "tirets simples" (simple dashes) may refer to Haar or Hadamard like vectors. Less structured grids (see Any grid) are also present, as in this "Répartition aléatoire de 40 000 carrés":

The following "3 doubles trames 0°, 30°, 60°" is far better structured. Connections. Or Connexions (cnx.org), at Rice University, providing Content Commons of free, open-licensed educational materials in fields such as music, electrical engineering and psychology. Remember Rice also hosts a Compressive Sensing repository.

RIP scene sends Smog (Noiselet today)

Recently, Processing Vein Mess was proposed as an applicative anagram for Compressive sensing. Its Compressed sensing avatar possesses a much nicer anagram: "RIP scene sends Smog". This was illustrated in a recent talk by Albert Cohen around the Restricted Isometry Property (RIP). For practical applications, Compressive sensing onset lies in the developpement of specific observation matrices. Noiselets are an example of those. Let us quote artist Michael Thieke about them:

Very sparse. Very minimal.
These musicians with their instruments make sounds that may not have been intended by the original inventors. They do this in a way that at first seems to be a very random. After a longer listen, the inspirations soak through. These “noiselets and sounduals” (my words entirely) may be improvised, but they are very potent in their expressive capability.

Real, or imaginery (for noiselet Matlab code)?

RIP, not RIP - a little bit on compressed sensing

Shall the Restricted Isometry Property (RIP) Rest In Peace (RIP again)?

This afternoon, Albert Cohen presented recent findings (made with Wolfgang Dahmen and Ronald DeVore) on Compressed sensing and best k-term approximation. These findings pertain to the fields of compressive sensing (CS, see Nuit Blanche for some live coverage) and non-linear approximation (NLA), namely: how do fixed linear measurements (CS) and adaptive linear measurements (NLA) compare for signal approximation in the lp norm? Both in a deterministic and a probabilistic fashion. A. Cohen nicely reviewed some references on the topic, including the interesting FRI (Finite Rate of Innovation) by Per Luigi Dragotti and Martin Vetterli (et al.), nicely covered here (we come back later on this approach) and the compressive sensing resources. He emphasized an approach based on decoupling (1) the information quantity and (2) the algorithms. The basic question is: given some unknown x, presumably sparse, linearly observed via y = Px, find a decoder D (generally non-linear) such that D(Px) is close enough to x. (P,D) is characterized by some "instance optimality" property. The work focuses on the characterization of the null space ker(P), since it essentially measures P's ambiguity in recovering a sparse x from y. The existence of the decoder is related (Lemma 3.1 from Compressed sensing and best k-term approximation) to the injectivity of P and to the relative spread of ker(P). Compared to Candès-Tao-Romberg Restricted Isometry Property, this approach differs both in the algorithmic approach and in its stability. It is important to mention, that the "ker" (null space) point of view still applies when x = Ms is a quasi-sparsified (through a wavelet transform for instance) version of some signal s, while PM would not satisfy RIP, P would.

Coming back to Finite Rate of Innovation. A signal band-limited in [-1/2 1/2] is perfectly reconstructed with at least one uniform measurement per period of time, through a cardinal sine kernel. Which may be considered as an innovation rate per time unit. While the Nyquist-Shannon condition is sufficient, it is not necessary. FRI derives conditions, based on parametric hypotheses on the signals (e.g. formed with p diracs per time period), for perfectly reconstructing signals after an acquisition made of a sampling kernel followed by uniform sampling. Interestingly, in contrast to standard CS, the sparsity in FRI resides in a continuous domain, and not in a discrete one. This seems somewhat more natural. On the other hand, non-uniform sampling is more efficiently dealt with CS.

For recent accounts on FRI:
Estimating Signals with Finite Rate of Innovation from Noisy Samples: A Stochastic Algorithm, by Vincent Y. F. Tan, Vivek K. Goyal
Compressive Sampling: Sparse Sampling of Signal Innovations, by Thierry Blu, Pier-Luigi Dragotti, Martin Vetterli, Pina Marziliano, Lionel Coulot
Nuit Blanche of the above paper
Nuit Blanche in Neutron transport (very cold neutrons?)

Ever Alone Hunk - Discretization and Orthogonality

Far far far away from the last post on compressive sensing, two other talks closer to every day problems with standard data (orthogonal matrices and quantization), which i have been watching during breakfast (quite a long one).

Ofer Zeitouni: A Correlation Inequality for Nonlinear Reconstruction, on the question whether the Karhunen-Loève transform remains optimal in the non-linear approximation for gaussian vectors. Karhunen-Loève transform (or Proper orthogonal decomposition, or Principal component analysis) is a central tool in statistics (often named after Hotelling) and signal processing, yielding "optimal" orthogonal transforms for uncorrelated data. The duet formed by Kari Karhunen and Michel Loève (even a triplet which Harold Hotelling) reassembles to an Ever Alone Hunk.

Zhidong Bai: Statistical Analysis for Rounding Data. Most of the discrete versions of the continuous setting deal with the discretization of time or space, i.e. sampling. Now what happens to the discrete amplitudes, i.e. discretization in the "value" domain? Especially that standard estimation of mean and variance are not consistant. Some counter-intuitive experiments with respect to the central limit theorem. Here, the larger, the worse.

Compressive sensing - Processing Vein Mess

This first short note on Compressive Sensing is, of course, anagram-driven since it has been applied for some Vein Mess Processing (see the images on the last page of Sparse MRI: The Application of Compressed Sensing for Rapid MR Imaging, by Michael Lustig, David Donoho and John M. Pauly). Some matlab code for compressed sensing MRI is made available. It is also dedicated to Igor Carron who carefully blogs and extracts information bits from the overwhelming litterature on CS (see Compressive Sensing Resources). For short, compressed or compressive sensing aims at saving bits by sampling and compressing structured signals at the same time, allowing some potential error. It is strongly related to sparse decompositions. An open challenge: find the sparsest/most compressed, but also most elegant, definition for "compressive sensing"! I do consider the four-word Wedderburn theorem "every finite field commutates" (in French: "tout corps fini commute", with open poetic interpretations) one of the beautiful mantras (sure, allowing an amount of metonymia, five-word versions are more correct). The right-side image of a "corps fini" is borrowed from here.

It is also meant to provide recordings from talks around the topic, for the unfortunate who could not attend. Most of the talks rest of the wonderful Diffusion des savoirs de l'Ecole Normale Supérieure site, especially for the Mathematical Foundations of Learning Theory conference.

• By Yves Meyer (in french): une avancée étonnante en imagerie numérique : le "Compressed Sensing"
• By Emmanuel Candes: The Dantzig Selector: Statistical Estimation when p is Larger than n
• By Vladimir Koltchinskii: Sparsity in High-Dimensional Learning Problems

Compressive sensing may solve some problems. Hope it could also contribute to stock exchange troubles by improving SEC Senses. While standard banks fail, filter banks (and wavelets of course) never deceive.

Israel - La ré si

Du fait de son invité cette année, le salon du livre ne pouvait que générer un post de L. Duv. on Israël. Cette tentative capilotractée ne dissimule pas que les lettres d'"israel" sont dans celles du "salon du livre". Israël, pays du livre et des livres, fait débat. En ajoutant l'article, le salon du livre devient Israel, null dove, si l'on insiste sur les malheurs de la colombe, ou Israel, dull oven, pour son soleil harassant. Si Israel - is real est un slogan tentant, on peut reprocher à cet état qu'il rase les maisons de certains de ses habitants. Mais le raïs est aussi son anagramme. Ce n'est pas étonnant que les anagrammes sèment le trouble dans cette ambiance de gématrie (qui associe chiffres et lettres et en tire des interprétations symboliques). En espérant qu'on parle enfin en paix (www.parlerenpaix.org) dans ce serial de lutte et d'incompréhension, écoutons un peu la radio (France-Culture) sur l'actualité du salon du livre et son invité, et laissons parler les écrivains. Ça commence par Amos Oz, avec Clémence Boulouque, dans À voie nue, sur 5 émissions :

Lecteur – mon ami, mon frère – ne t’y trompe pas. Metsoudat-Ram n’est pas le miroir du monde. Ses habitants ne sont pas d’ici. Le village est le reflet d’une terre lointaine, perdue dans la brume. Ailleurs, peut-être.

(extrait de Ailleurs)

Sari Nusseibeh, Des clés pour la Palestine, suivra à partir du lundi 17 mars, dans la même émission. Une série de dix émissions "En hébreu dans le texte" ou pages arrachées à la littérature israélienne sont diffusées à partir du 10 mars, et offre un parcours dans le temps du 19e au 21e siècle. Samedi 15 mars sont également diffusées en fictions des pages arrachées à Amos Oz (à 20h) et les nouvelles voix de la littérature israélienne (à 22h10).
Régis Debray intervient sur le thème Religion et état, pour son Candide en terre sainte, un parcours par delà les deux rives du Jourdain, qu'on ne peut plus traverser sans le savoir. Autant en garder seulement quelles notes pour la route : la si ré (le désert n'est pas un sol fa si...)

Les portails :
Émissions enregistrées en publics
Autour d'Israel
Diaporama sonore, avec Claude Vigée

Lazare Ponticelli : le der des der

Lazare Ponticelli, le dernier des poilus : la 1ere guerre mondiale a rendu son lot capillaire zen. Une dernière tranchée à sa taille se referme sur le dernier souvenir vivant d'une époque de bruit et de fureur. Né italien, menant la charge dès 16 ans pour un pays, pas encore le sien, où les immigrés sont trop souvent considérés comme charge. Il en subsiste sur la toile au moins quatre épisodes sur LibéLabo, tant que restera la mémoire numérique...

Le compte-rendu de Maître Eolas
Le compte-rendu de Wiki

Wavelet activity - ondelette, et dolente ?

Fortunately, the wavelet activity is not as sparse as the representation it provides. Indeed, there is no such thing as the lazy wavelet (l'ondelette indolente). Recent additions from the wavelet world, namely wavelet names in "*let" include treelets, noiselets, needlets and a connection to Vassili Strela multiwavelet toolbox, which was long unavailable. Wavelets do not suffer (inactivity). The anagramatic connection linking "ondelette" "et dolente ?" is thus... wrong. There is such thing as the lazy wavelet. Contradiction? No, we encourage star-let submission!

Un test, qui sait ? (Des statistiques...)

Pénombre offre un espace public de réflexion et d'échange sur l'usage du nombre dans les débats de société. On sait, après Disraeli (premier ministre britannique au XIXe siècle) qu'il existe trois sortes de mensonges : les petits mensonges, les gros mensonges et les statistiques ! Pourtant, bien employée, la statistique mets en lumière les vérités profondes du monde. Par exemple cette découverte étonnante dans le numéro 42 de la lettre de pénombre de la constance de la population mondiale par personne.

Cette statistique peut-être mise en rapport avec cette autre découverte de l'église du monstre spaghetti volant (Church of the Flying Spaghetti Monster). Comme quoi, les structures religieuses peuvent également contribuer à la science. Et donc, on y apprend l'incidence découverte de la diminution du nombre de pirates sur le réchauffement global. En parallèle à la réduction du duo production/consommation, à la mise en œuvre de solutions de stockage du CO2, on attend des initiatives pour lutter contre le déclin de la flibuste, des corsaires, de la boucane :
En ce moment, les statistiques prennent une importance particulière dans mes préoccupations de recherche. Avec par exemple l'usage du principe de Stein (ou le paradoxe, le phénomène). Il indique, pour le dire vite, que si l'on estime trois paramètres (ou plus) simultanément, leur estimation combinée est plus précise que si on les estime séparément. Cela conduit à l'estimateur de James-Stein. Récemment, ce principe a été adapté pour estimer des signaux et des images à partir de leurs versions bruitées à partir de leurs coefficients de décomposition en ondelettes, par exemple avec des décompositions en ondelettes en arbre dual (ou multirésolution en général), pour des données multicanaux (cf. http://arxiv.org/abs/0712.2317).

Et plus récemment les statistiques robustes connaissent un regain d'intérêt. À suivre...

Quentin Garel, en quel granit (en bords de Seine)

En quel granit furent taillés les géants qui marchaient sur le globe ? De quel bois les oiseaux ? Quentin Garel offre à l'œil quelques vestiges du temps passé, quelques résidus de mort à venir.

Becs et crânes d'oiseaux, tête de canard décharnée, mandibules de porc, bois ou métal. Ces œuvres d'art, quel engin les a déposées sur les rives de la Seine, dans le parc du Chemin de l'Île ? À Nanterre, 14 hectares en bord de Seine, entre A14, friches industrielles, décharges et centre de recyclages. Où ces rognures de l'industrie consommante se trouvent fort bien. Quelques indications pour la visite du parc et les infos officielles. Le dimanche 30 mars est prévue une visite guidée de parc. En savoir plus sur Quentin Garel ? À la galerie L.J. Beaubourg, les photos de Christelle (sur le blog de Buzenval à Miromesnil), à la Galerie Forni (Bologna, Italie). Dans Le bestiaire de Quentin Garel (fiche technique), documentaire (Mirage Illimité) de 2006 de Thomas Lallier, en co-réalisation avec Alyssa et Myrha Verbizh.

Quelques autres photos, de ci, de là :

http://chrislgn9.canalblog.com/images/garel1.jpg
http://chrislgn9.canalblog.com/images/garel7.jpg

http://chrislgn9.canalblog.com/images/garel6.jpg
http://chrislgn9.canalblog.com/images/garel4.jpg
http://chrislgn9.canalblog.com/images/garel2.jpg
http://chrislgn9.canalblog.com/images/t-garel5.jpg
http://chrislgn9.canalblog.com/images/t-garel3.jpg
http://www.galleriaforni.it/artistiinesclusiva/garel.jpg
http://www.exibart.com/profilo/imgpost/rev/856/rev22856(1)-ori.jpg

Le Moscar des pickpockets

On en avait parlé ici même, et sur les enquêtes en stand-by d'Oslo Films, du Mozart des pickpockets. Une petite perle de court-métrage, drôle et bien écrit, de Philippe Pollet-Villard, et avec, et avec Richard Morgièvre, compère écrivain. Très primé par chez nous, il vient de recevoir l'Oscar 2008 du meilleur court-métrage de fiction.

Ce qu’il y a de bien c’est qu’on saura jamais s’il est intelligent ce gamin. Ça doit être ça l’avantage d’être sourd et muet, t’es pas dans la compétition

Ce court est disponible en DVD (édité par F-CAN), avec un autre court-métrage, La baguette, dont les sub-héros sont les mêmes Philippe et Richard, larcineurs à la petite semaine. Bon investissement, il se revoit facilement.

Actualités du court-métrage - Clermont-Ferrand

L'actualité est au festival international du court métrage du Clermont-Ferrand

Une diffusion des films courts :
MK2, soirée courts métrages n°96
Mardi 12 février à 20H30 et 22h30 (Mk2 quai de Seine, métro : Jaurès, Stalingrad)

Les prix du festival :
Compétition nationale : Como todo el mundo, de Franco Lolli
Compétition internationale : Auf der Strecke, de Reto Caffi
Compétition labo : Pikapika, Lightning Doodle Project, de Kazue Monno, Takeshi Nagata
Tout le palmarès : sur le site du festival

À la radio :
Tout arrive (France Culture) du mercredi 6 février 2008 (partie 1, partie 2 en audio)
Avec un focus sur Gravity de Nicolas Provost (synthèse d'innombrables baisers hollywoodiens), et Stone people de Leonid Rybakov

Et les prochains festivals :
Festivals en France
Festivals à l'étranger

Microfilm - I, Film Corm

Un excellent concert de Microfilm hier soir 8 février 2008 au Glaz'Art (www.glazart.com). Avec l'insigne avantage d'être passé en première partie, permettant de quitter la salle au milieu (voire au début) des jérémiades bruitistes d'Arnaud Michniak. Un anagramme de Microfilm ? I, Film Corm ("ma vie de bulbe cinématographique") donne une petite idée de ce groupe surtout instrumental, dont germent des samples de paroles et quelques scènes retouchées de vieux films (comme par un filtre médian), projetées sur un écran. Comme pour Secret camera, tiré de leur premier album A journey to the 75th, et dont le clip semble extrait de Journey To The Far Side Of The Sun (Doppleganger).

Le microfilm est/était un support de stockage analogique reproduisant les pages de livres ou périodiques, des documents juridiques et des dessins techniques. ...

selon la source wikipedia usuelle. Microfilm, c'est désormais :

un groupe de 4 musiciens de Poitiers, teinté de post-rock, distingué par un ep éponyme en 2003, suivi par A journey to the 75th en 2004, puis Stereodrama (livré avec lunettes de relief bleu et rouge) en 2007, et qui concerte en Chine.

Depuis leur chanson André, ils sont devenus avec Stereodrama des Atom readers avec la désintégration de leur chatte Dandinette ( par ailleurs, une technique de pêche) :

• Lui : C'est maintenant un flot d'atomes félins.
• Elle : Mais où est-elle allée ?
• Lui : Dans l'espace...

Ce concert énergique a diffusé Tout l'univers, La fille qui en savait trop, Dpt. 4, Johnny and Barbara, Secret Camera, The Night of J&B (une référence surement à la vieille publicité pour du whisky : J&B pour la nuit), Dpt. 1, Death race 68, Fancy Nancy, Margaret on the rocks.

Plus d'infos, et pour les écouter :
www.microfilm.tv
www.myspace.com/microfilmband
Pour les voir, les dates :
www.microfilm.tv/dates.html
MySpace : dates

Pour finir, les mots du batteur préparant un final à la caisse claire au centre de la salle : "on ne parle pas, mais ça fait du bien de crier quand même".

Nota : il existe un autre Microfilm (www.microfilmmusic.com), groupe techno de Matthew Mercer et Matt Keppel, fondé en 2006. Et c'est pas bien de faire des comparaisons, mais Microfilm me redonne l'envie d'écouter The God Machine, groupe saturé défunt du début des années 1990, avec deux excellents albums (eux aussi) : Scenes From The Second Storey et One Last Laugh In A Place Of Dying. (quelques infos supplémentaires par ici et ici). Allez, une petite vidéo interview live de Microfilm pour la route :

Microfilm, interview TV + morceaux live par le webzine sourdoreille

Sans corps

Philippe Khorsand (fiche IMDB), pensionnaire de la Comédie Française, est décédé le 29 janvier 2008.

Révélé à la télévision par la série Palace de Jean-Michel Ribes, il a commencé à jouer au cinéma pour Georges Launter (« Laisse aller, c'est une valse »), le rôle d'Antoine dans « Mes meilleurs copains » de Jean-Marie Poiré, celui de Bergosa dans le trop-vite-dénigré « Total western » de Éric Rochant (avec un Kalfon furioso) et, il faut bien le dire, une série de publicité pour une assurance cotonneuse. Fama crescit eundo, la rumeur croît en s'éloignant, et donne dans le désordre MAAF : escroc de nuit. Son dernier rôle est pour Jean-Michel Ribes, dans « Musée haut, musée bas », prévu cette année.

Appelez-moi le directeur ! Trop tard, le patron l'a rappelé ! Il l'a eu un jour... Il l'a eu !

"Le musée est-il chaud, est-il froid ? Est-il haut, est-il bas ? Y a-t-il des veaux, y a t-il des rats ?"

Propriétaires de tous pays, unissez-vous

À la rentrée dernière, je proposais une lecture pudique de la lettre de Guy Hoquet. Il manquait le geste à la parole. Voila qui est FAI.

Il joue du piano. Debout

Daniel Barenboim vient (samedi) de recevoir un passeport palestinien. C'est un acte fort pour la coexistence israelo-palestienne. Daniel Barenboim avait joué régulièrement en territoire palestinien, il avait également fait un symbole de son West-Eastern Diwan Orchestra, ou Orchestre Divan oriental-occidental (avec Edward Said), fondé en 1999 en référence à un recueil de poèmes de Goethe (pour son 250e anniversaire), comprenant notamment des musiciens israéliens et de pays arabe (Palestine, Égypte, Syrie). C'est peut-être un détail pour vous, comme l'initiative Parler en paix, mais pour moi...

La psilopa et un dessin de Pessin

La psilopa (petrolei) est une mouche dont la larve vit à proximité des mares de pétrole, et dont j'oubliais régulièrement le petit nom, avant d'en faire un titre. Du pétrole à la mouche, lien direct avec ce dessin de Pessin. Il est loin le temps où on ramassait le pétrole à la louche (hormis en période de marée noire).

À 100 dollars le baril, ça reste moins cher que le caviar (c) Pessin

À grimper, le baril fera peut-être un jour de la voiture l'équivalent des imprimantes ou des rasoirs à x lames d'aujourd'hui : un produit d'appel qui ne coûte rien et dont le fonctionnement requiert l'achat régulier d'annexes hors de prix (les fameuses lames ou les cartouches d'encre).


La Tata Nano sera vendue en Inde aux alentours de 2200 dollars (environ 1 700 euros, mais ça ne va pas durer). Coffre avant pour une petite valise, compteur kilométrique analogique, boîte de vitesses issue d'un système inventé par Léonard de Vinci, elle consommerait 5 litres aux 100 kilomètres. Aux tarifs d'ici, elle aurait consommé son prix en essence en 25 000 km. Kilométrage réalisé en moyenne entre 1 et 2 ans. À quand la mobylette au caviar ?

Un anagramme quand même ? Psilopa petrolei = Stopperai le poil. Rien à voir.