BLOG    
 

Self-similarity analysis: a selected bibilography

Recently there has been much interest in using similarity matrices to analyze linear media (audio, video, text, sequences). A similarity matrix is a 2D characterization of all pairwise similarity measurements. Variants are also called "recurrence plots" or "dotplots." Though similar in some respects to graph-theoretic adjacency matrices, the latter are sufficiently different (no explicit dimension on the diagonal, so permutations don't matter) that we don't consider them here. As usual, it it likely we are missing important contributions: if you know of any we have missed, please contact us and let us know.

The earliest references seem to appear in the bioinformatics literature:


Maizel, J. V.  Jr, Lenk, R. P.,  "Enhanced graphic matrix analysis of nucleic acid and  protein sequences." In Proc. National Acadamy of Science, USA. Vol. 78 No. 12., pp 7665-9. Dec 1981 (Abstract)

Shapiro B., Nussinov R., Lipkin L., Maizel J. Jr., "An interactive dot matrix system for locating potentially significant features in nucleic acid molecules." Journal of Biomolecular Structure & Dynamics Vol 4., No. 5, pp. 697-706. April 1987 (Abstract)

Around the same time, a variant called "recurrence plots" were explored in the mathematical physics community to investigate 1-D recurrent and chaotic systems.


Eckman, J.P., et. al., “Recurrence Plots of Dynamical Systems,” in Europhys. Lett. Vol. 4, No 973,  November 1987

Marwan, Norbert, "An introduction to recurrence plots, cross recurrence plots and recurrence quantification analysis." http://www.recurrence-plot.tk (also has a thorough bibliography)

Church and Helfman also used a similarity-matrix approach for visualizing the structure of large text corpora:


Church, K.W., and Helfman, J.I. "Dotplot: A Program for Exploring Self-Similarity in Millions of Lines of Text and Code," Journal of Computational and Graphical Statistics, Vol. 2 No. 2, pp. 153-174, June 1993  Examples: http://imagebeat.com/dotplot/

Cutler and Davis used a matrix approach to detect periodic motion in video:


Cutler, R. and Davis, L. “Robust Real-Time Periodic Motion Detection, Analysis, and Applications,” IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), Vol. 22 No. 8, pp. 781-796,  August 2000

A curious paper used "linkplot" (adjacency) matrices to visualize hypertext link structure. Note that this predates the WWW by several years!


Bernstein, M., J.D. Bolter, M. Joyce and E. Mylonas (1991), "Architectures for Volatile Hypertext," in Proc. Hypertext '91: Third ACM Conference. on Hypertext, ACM, 243-260. http://portal.acm.org/citation.cfm?p=122999&coll=portal&dl=ACM


Recently, many researchers have used matrix approaches to analyze the structure of audio:


Bartsch, M. and Wakefield, G.H., "To Catch a Chorus: Using Chroma-Based Representations For Audio Thumbnailing." Proc. Workshop on Applications of Signal Processing to Audio and Acoustics, WASPAA (2001), IEEE.
http://musen.engin.umich.edu/papers/bartsch_wakefield_waspaa01_final.pdf

Dannenberg, Roger, and Hu, ``Pattern Discovery Techniques for Music Audio,'' Proc. ISMIR 2002: Third International Conference on Music Information Retrieval, M. Fingerhut, ed., Paris: IRCAM, October 2002, pp. 63-70. Expanded to: Dannenberg and Hu, ``Pattern Discovery Techniques for Music Audio,'' Journal of New Music Research, (to appear).
http://www-2.cs.cmu.edu/afs/cs.cmu.edu/user/rbd/www/papers/ismir2002pattern.pdf

Geoffroy Peeters, Amaury La Burthe and Xavier Rodet. "Toward Automatic Music Audio Summary Generation from Signal Analysis." Proc. ISMIR 2002: Third International Conference on Music Information Retrieval, M. Fingerhut, ed., Paris: IRCAM, October 2002.
http://ismir2002.ismir.net/proceedings/02-FP03-3.pdf

J. Wellhausen and H. Crysandt,  "Temporal Audio Segmentation Using MPEG-7 Descriptors." Proc. SPIE Storage and Retrieval for Media Databases, SPIE Vol. 5021, Santa Clara (CA), USA , January 2003
http://www.ient.rwth-aachen.de/team/wellhausen/publications/ei2003/EI2003_5021_41.pdf

Masataka Goto: A Chorus-Section Detecting Method for Musical Audio Signals, Proc. 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2003), Vol. V, pp. 437-440, April 2003. http://staff.aist.go.jp/m.goto/PAPER/ICASSP2003goto.pdf

Chai, Wei and Vercoe, Barry. "Structural Analysis Of Musical Signals For Indexing and Thumbnailing."  Proc. of ACM/IEEE Joint Conference on Digital Libraries, May 2003.  http://web.media.mit.edu/~chaiwei/papers/chaiw_structure.pdf

Lie Lu, Stan Li,  Wen-Yin Liu, Hong-Jiang Zhang.  " Audio Textures". Proc. of IEEE International Conference on Acoustics, Speech and Signal Processing 2002 (ICASSP02), Vol. II pp 1761-1764.
http://www.research.microsoft.com/asia/dload_files/group/mc/2004/1-TSAP_AudioTexture.pdf

Michael Casey, http://musicstructure.com


At FXPAL, we have investigated self-similarity analysis on audio, video and time indexes. Here is a list of our relevant publications.

Copyright ©1999-2010 FX Palo Alto Laboratory | Send feedback to the webmaster