Continuous wavelets of compact support alpha can be built,[1] which are related to the beta distribution. The process is derived from probability distributions using blur derivative. These new wavelets have just one cycle, so they are termed unicycle wavelets. They can be viewed as a soft variety of Haar wavelets whose shape is fine-tuned by two parameters and . Closed-form expressions for beta wavelets and scale functions as well as their spectra are derived. Their importance is due to the Central Limit Theorem by Gnedenko and Kolmogorov applied for compactly supported signals.[2]

Beta distribution

edit

The beta distribution is a continuous probability distribution defined over the interval  . It is characterised by a couple of parameters, namely   and   according to:

 .

The normalising factor is  ,

where   is the generalised factorial function of Euler and   is the Beta function.[3]

Gnedenko-Kolmogorov central limit theorem revisited

edit

Let   be a probability density of the random variable  ,   i.e.

 ,   and  .

Suppose that all variables are independent.

The mean and the variance of a given random variable   are, respectively

   .

The mean and variance of   are therefore   and  .

The density   of the random variable corresponding to the sum   is given by the

Central Limit Theorem for distributions of compact support (Gnedenko and Kolmogorov).[2]

Let   be distributions such that  .

Let  , and  .

Without loss of generality assume that   and  .

The random variable   holds, as  ,    

where   and  

Beta wavelets

edit

Since   is unimodal, the wavelet generated by

  has only one-cycle (a negative half-cycle and a positive half-cycle).

The main features of beta wavelets of parameters   and   are:

 

 

The parameter   is referred to as “cyclic balance”, and is defined as the ratio between the lengths of the causal and non-causal piece of the wavelet. The instant of transition   from the first to the second half cycle is given by

 

The (unimodal) scale function associated with the wavelets is given by

   .

A closed-form expression for first-order beta wavelets can easily be derived. Within their support,

 

 
Figure. Unicyclic beta scale function and wavelet for different parameters: a)  ,   b)  ,   c)  ,  .

Beta wavelet spectrum

edit

The beta wavelet spectrum can be derived in terms of the Kummer hypergeometric function.[4]

Let   denote the Fourier transform pair associated with the wavelet.

This spectrum is also denoted by   for short. It can be proved by applying properties of the Fourier transform that

 

where  .

Only symmetrical   cases have zeroes in the spectrum. A few asymmetric   beta wavelets are shown in Fig. Inquisitively, they are parameter-symmetrical in the sense that they hold  

Higher derivatives may also generate further beta wavelets. Higher order beta wavelets are defined by  

This is henceforth referred to as an  -order beta wavelet. They exist for order  . After some algebraic handling, their closed-form expression can be found:

 

 
Figure. Magnitude of the spectrum   of beta wavelets,     for Symmetric beta wavelet  ,  ,  
 
Figure. Magnitude of the spectrum   of beta wavelets,     for: Asymmetric beta wavelet  ,  ,  ,  .

Application

edit

Wavelet theory is applicable to several subjects. All wavelet transforms may be considered forms of time-frequency representation for continuous-time (analog) signals and so are related to harmonic analysis. Almost all practically useful discrete wavelet transforms use discrete-time filter banks. Similarly, Beta wavelet[1][5] and its derivative are utilized in several real-time engineering applications such as image compression,[5] bio-medical signal compression,[6][7] image recognition [9][8] etc.

References

edit
  1. ^ a b de Oliveira, Hélio Magalhães; Schmidt, Giovanna Angelis (2005). "Compactly Supported One-cyclic Wavelets Derived from Beta Distributions". Journal of Communication and Information Systems. 20 (3): 27–33. arXiv:1502.02166. doi:10.14209/jcis.2005.17.
  2. ^ a b Gnedenko, Boris Vladimirovich; Kolmogorov, Andrey (1954). Limit Distributions for Sums of Independent Random Variables. Reading, Ma: Addison-Wesley.
  3. ^ Davis, Philip J. (1968). "Gamma Function and Related Functions". In Abramowitz, Milton; Stegun, Irene (eds.). Handbook of Mathematical Functions. New York: Dover. pp. 253–294. ISBN 0-486-61272-4.
  4. ^ Slater, Lucy Joan (1968). "Confluent Hypergeometric Function". In Abramowitz, Milton; Stegun, Irene (eds.). Handbook of Mathematical Functions. New York: Dover. pp. 503–536. ISBN 0-486-61272-4.
  5. ^ a b Ben Amar, Chokri; Zaied, Mourad; Alimi, Adel M. (2005). "Beta wavelets. Synthesis and application to lossy image compression". Advances in Engineering Software. 36 (7). Elsevier: 459–474. doi:10.1016/j.advengsoft.2005.01.013.
  6. ^ Kumar, Ranjeet; Kumar, Anil; Pandey, Rajesh K. (2012). "Electrocardiogram Signal compression Using Beta Wavelets". Journal of Mathematical Modelling and Algorithms. 11 (3). Springer Verlag: 235–248. doi:10.1007/s10852-012-9181-9. S2CID 4667379.
  7. ^ Kumar, Ranjeet; Kumar, Anil; Pandey, Rajesh K. (2013). "Beta wavelet based ECG signal compression using lossless encoding with modified thresholding". Computers & Electrical Engineering. 39 (1). Elsevier: 130–140. doi:10.1016/j.compeleceng.2012.04.008.
  8. ^ Zaied, Mourad; Jemai, Olfa; Ben Amar, Chokri (2008). "Training of the Beta wavelet networks by the frames theory: Application to face recognition". 2008 First Workshops on Image Processing Theory, Tools and Applications. IEEE. pp. 1–6. doi:10.1109/IPTA.2008.4743756. eISSN 2154-512X. ISBN 978-1-4244-3321-6. ISSN 2154-5111. S2CID 12230926.

Further reading

edit
  • W.B. Davenport, Probability and Random Processes, McGraw-Hill, Kogakusha, Tokyo, 1970.
edit