Sistema de evaluaci

1

2
(No Transcript)
3
STFT (Short time Fourier transform)
Or windowed Fourier transform
4
(No Transcript)
5
This is the basis for most computer generated
spectrograms (display pixel intensity on a log
scale by limiting the dynamic range to about
60-80 dB).
6
Ejemplos de espectrogramas
Here is the sum of two parallel linear chirps
with its spectrogram.
7
Here is the sum of two hyperbolic chirps and its
spectrogram.
8

• Introducción a Wavelets

9
four frequency components at different times.
The interval 0 to 250 ms is a sinusoid of 300 Hz,
and the other 250 ms intervals are sinusoids of
200 Hz, 100 Hz, and 50 Hz
w(t)exp(-a(t2)/2)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
Wavelet de Morlet
24
The mexican hat wavelet
25
                                                Gráficos de varios tipos distintos de wavelets. (a) Wavelet de Haar, (b) Wavelet de Daubechies, (c) Wavelet de Morlet. (Cortesía de Ofer Levi, Universidad de Stanford)
26
Escala
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(CWT) Continuous Wavelet Transform
escalograma
These signals are drawn from a database signals
that includes event related potentials of normal
people, and patients with Alzheimer's disease.
33
(No Transcript)
34
En un espectrograma
35
En un escalograma
36
Suma de dos señales CHIRP hiperbólicas
Windowed fourier transform (Espectrograma)
Continuous Wavelet Transform CWT (Escalograma)
37

• Representación frecuencia-tiempo para
• Datos muestreados
• (b) FT
• (c) WFT
• (d) CWT

38
(No Transcript)
39
MATLAB cwt
COEFS cwt(S,SCALES,'wname') computes the
continuous wavelet coefficients of the vector S
at real, positive SCALES, using the wavelet whose
name is 'wname'
40
(No Transcript)
41
DWT (Discrete Wavelet Transform)
42
Ahora dejamos fija la Ondeleta y lo que vamos
comprimiendo por etapas es la señal
43

• El análisis multiresolución se consigue a través
  de filtrado y submuestreo de la señal original.
• La exploración en tiempo se consigue a través de
  operaciones de convolución (filtrado digital).

44
Sub-band coding
45
Sub-band coding algorithm
46
(No Transcript)
47
Transformada inversa
48
(No Transcript)
49
2-D Discrete Wavelet Transform
50
(No Transcript)
51
(No Transcript)
52
(No Transcript)
53
(No Transcript)
54
(No Transcript)
55
(No Transcript)
56
Wavelet Packet
57
(No Transcript)
58
En resumen
CWT
DWT
2D - DWT
59
... En resumen
CWT
DWT
2D - DWT
60

• http//www.gisdevelopment.net/technology/ic/techip
  0003a.htm

COMPRESION DE LA DCT A WAVELETS
http//www.acm.org/crossroads/xrds6-3/sahaimgcodin
g.htmlFig6
61
(No Transcript)