Download presentation
Presentation is loading. Please wait.
Published byMorris Collins Modified over 9 years ago
1
Image Transforms Instructed by : J. Shanbezadeh Email : Shanbehzadeh@gmail.com 1Jamshid Shanbehzadeh
2
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –B–Basis Image (m 1,m 2 ) –R–Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2–2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –F–Forward Separable Transform –I–Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 2Jamshid Shanbehzadeh
3
Applications of Image Transforms Extracting Features from Images –In Fourier Transform, the average dc term is proportional to the average image amplitude Image Compression –Dimensionality Reduction 3Jamshid Shanbehzadeh
4
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 4Jamshid Shanbehzadeh
5
Types of Image Transforms Unitary Transforms Fourier Transforms Cosine, Sine, Hartley Transforms Hadamard, Haar Wavelet Transforms Ridglet, Curvelet, Contourlet 5Jamshid Shanbehzadeh
6
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 6Jamshid Shanbehzadeh
7
F(0,0)=f(0,0).A(0,0,0,0)+f(0,1)A(0,1,0,0)+f(0,2)A(0,2,0,0)+….+f(0,N2-1)A(0,N2-1,0,0) 2-D Transforms : the Forward Transform Kernel Forward transform of the N 1 *N 2 image array F(n 1,n 2 ) : به ازای هر m 1 و m 2 یک تصویر پایه ساخته می شود. n 1 و n 2 پیکسلهای تصویر در فضای جدید هستند. 7Jamshid Shanbehzadeh
8
Basis Image (m 1,m 2 ) 8Jamshid Shanbehzadeh
9
Basis Image (m 1,m 2 ) 9 Jamshid Shanbehzadeh
10
10Jamshid Shanbehzadeh
11
پیکسلهای تصویر اصلی را در پیکسلهای تصویر پایه، نظیر به نظیر در یکدیگر ضرب داخلی می نماییم. 11Jamshid Shanbehzadeh
12
یعنی کل تصویر را پیمایش می نمایند. مقایسه تبدیل یک بعدی و دوبعدی 12Jamshid Shanbehzadeh
13
Reverse 2-D Transforms A reverse or inverse transformation provides a mapping from the transform domain to the image space as given by : B(n 1,n 2 ; m 1,m 2 ) : the Inverse Transform Kernel کرنل مورد استفاده در تبدیل تصاویر بایستی معکوس پذیر باشد. 13Jamshid Shanbehzadeh
14
Basis Inverse Transform Image (m 1,m 2 ) 14Jamshid Shanbehzadeh
15
به ازای هر n1 و n2 یک تصویر پایه ساخته میشود، اگر تصاویر پایه بر هم عمود باشند. 15Jamshid Shanbehzadeh
16
2-D Unitary Transforms The transformation is unitary if the following orthonormality conditions are met: 16Jamshid Shanbehzadeh
17
Inner Product 17Jamshid Shanbehzadeh
18
Inner Product 18Jamshid Shanbehzadeh
19
Image Size(IS) =512 X 512 Number of Operations = IS X IS(Mul)+(IS X IS-1) (Addition) for one element =512 X 512(Mul) +(512 X 512 -1)(Addition) Number of operations for all = 512 X 512 ( 512 X 512(Mul) +(512 X 512 -1)(Addition) ) ضرب داخلی تصاویر 19Jamshid Shanbehzadeh
20
Number Operations = 67,108,864(Multiplications)+1,032,192(additions) Image Size(IS) =512 X 512 Block Size(BS) =8 X 8 Number of Blocks(NB) =128 X 128 Size of Basis Image(SBI) =8 X 8 Number of Operations = NB X {BS X SBI(Mul)+(BS-1) (Addition)} =128 X 128{(64 X 64)Mult+63(Addition)} برای کاهش حجم محاسبات، تصاویر را به بلاکهایی تقسیم می نماییم: بلوک بندی تصاویر حجم محاسبات علیرغم کاهش، زیاد است. 20Jamshid Shanbehzadeh
21
16,384X(512(Mult)+448(additions))=8,388,608(Multi)+7,340,032(Additions) If we perform matrix multiplication, then we have for two N X N matrixes: Number of operations (NO)= N X N {N(Mul) + (N-1)(addition)} Number of Image Blocks (NIB) = Image Size/(NXN) Total Number of Operations(TNO)=NIB X NO Matrix multiplication حجم محاسبات بسیار کاهش می باشد. 21Jamshid Shanbehzadeh
22
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 22Jamshid Shanbehzadeh
23
Separable Transforms The transformation is said to be separable if its kernels can be written in the form Where the kernel subscripts indicate row and column one-dimensional transform operations. 23Jamshid Shanbehzadeh
24
A separable two-dimensional unitary transform can be computed in two steps: First, a one-dimensional transform is taken along each column of the image, yielding Next, a second one-dimensional unitary transform is taken along each row of P(m 1,m 2 ), giving Separable Transforms 24Jamshid Shanbehzadeh
25
Separable Transforms 25Jamshid Shanbehzadeh
26
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 26Jamshid Shanbehzadeh
27
Forward Transform F and f denote the matrix and vector representations of a signal array. F and f be the matrix and vector forms of the transformed signal. The two-dimensional unitary transform is given by F=Af Where A is the forward transformation matrix. 27Jamshid Shanbehzadeh
28
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 28Jamshid Shanbehzadeh
29
Reverse Transform The inverse transform is f = Bf B represents the inverse transformation matrix B = A -1 29Jamshid Shanbehzadeh
30
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 30Jamshid Shanbehzadeh
31
Unitary Matrix (Transform) For a unitary transformation, the matrix inverse is given by A -1 = A *T A is said to be a unitary matrix 31Jamshid Shanbehzadeh
32
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 32Jamshid Shanbehzadeh
33
Orthogonal Matrix (Transform) A real unitary matrix is called an orthogonal matrix. For such a matrix, A -1 = A T 33Jamshid Shanbehzadeh
34
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 34Jamshid Shanbehzadeh
35
Separable Transforms If the transform kernels are separable such that Where A R and A C are row and column unitary transform matrices. 35Jamshid Shanbehzadeh
36
The transformed image matrix can be obtained from the image matrix by Forward Separable Transforms F 36Jamshid Shanbehzadeh
37
Inverse Separable Transforms The inverse transformation is given by F = B C F B R T Where B C = A C -1 and B R = A R -1 37Jamshid Shanbehzadeh
38
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 38Jamshid Shanbehzadeh
39
Forward Fourier Transform کرنل دوبعدی جدایی پذیر 39Jamshid Shanbehzadeh
40
مقایسه تبدیل یک بعدی و دوبعدی 40Jamshid Shanbehzadeh
41
Inverse Fourier Transform Fourier Transform : Inverse Fourier Transform : 41Jamshid Shanbehzadeh
42
Fourier Transform (Separable) تبدیل دوبعدی را به صورت سینوسی و کسینوسی می نویسیم: 42Jamshid Shanbehzadeh
43
Fourier Transform (Separable) 43Jamshid Shanbehzadeh
44
Fourier transform basis functions, N=16 44Jamshid Shanbehzadeh
45
قسمت حقیقی مقادیر تصاویر پایه DFT قسمت موهومی تصاویر پایه DFT 45Jamshid Shanbehzadeh
46
تصویر اصلی اندازه تبدیل فوریه تصویر اصلی (مبدا به وسط انتقال یافته است.) اندازه تبدیل فوریه تصویر اصلی با استفاده از لگاریتم اندازه ها فاز تبدیل فوریه 46Jamshid Shanbehzadeh
47
تصویر اصلی تبدیل فوریه آن 47Jamshid Shanbehzadeh
48
دو نمونه تصویر اصلی تبدیل فوریه تصاویر چرخش یافته تصاویرتبدیل فوریه تصاویر حساسیت تبدیل فوریه به چرخش 48Jamshid Shanbehzadeh
49
Fourier Transform Properties The spectral component at the origin of the Fourier domain is equal to N times the spatial average of the image plane. 49Jamshid Shanbehzadeh
50
Zero-frequency term at the center Multiplying the image function by factor (-1) j+k یعنی با ضرب f(x,y) در (-1) x+y مبدا تبدیل فوریه f(x,y) به مرکز مربع فرکانسی N X N متناظرش انتقال داده می شود. 50Jamshid Shanbehzadeh
51
The Fourier transform in vector-space form : F = Af f = A* T F f and F are vectors obtained by column scanning the matrices f and F. F and f denote the matrix and vector representations of an image array. F and f be the matrix and vector forms of the transformed image. Fourier transform in vector-space 51Jamshid Shanbehzadeh
52
52Jamshid Shanbehzadeh
53
Fourier Transform Properties 53Jamshid Shanbehzadeh
54
Fourier Transform Properties Substitution u = - u and v = -v 54Jamshid Shanbehzadeh
55
55Jamshid Shanbehzadeh
56
a) Original imageb) Phase only image c) Contrast enhanced version of image (b) to show detail Phase data contains information about where objects are in the image Fourier Transform Phase Information 56Jamshid Shanbehzadeh
57
a) Original imageMagnitude of the Fourier spectrum of (a) Phase of the Fourier spectrum of (a) d) Original image shifted by 128 rows and 128 columns Magnitude of the Fourier spectrum of (d) Phase of the Fourier spectrum of (d) Translation Property با شیفت دادن تصاویر، در فاز حاصل از تبدیل فوریه تغییر ایجاد می شود اما اندازه آن ثابت خواهد ماند. 57Jamshid Shanbehzadeh
58
g) Original imageMagnitude of the Fourier spectrum of (g) Phase of the Fourier spectrum of (g) These images illustrate that when an image is translated, the phase changes, even though magnitude remains the same. Translation Property 58Jamshid Shanbehzadeh
59
a) Original image b) Fourier spectrum image of original image c) Original image rotated by 90 degrees d) Fourier spectrum image of rotated image Rotation results in Corresponding Rotations with Image and Spectrum Rotation Property 59Jamshid Shanbehzadeh
60
The test image has been scaled over unit range Where is the clipping Factor and is the maximum coefficient magnitude. 60Jamshid Shanbehzadeh
61
Another form of amplitude compression is to take the logarithm of each component as given by Where a and b are scaling constants 61Jamshid Shanbehzadeh
62
DIRECT REMAPCONTRAST ENHANCED LOG REMAP Cam.pgm An Ellipse Displaying DFT Spectrum with Various Remap Methods 62Jamshid Shanbehzadeh
63
DIRECT REMAPCONTRAST ENHANCED LOG REMAP House.pgm A Rectangle Displaying DFT Spectrum with Various Remap Methods 63Jamshid Shanbehzadeh
64
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 64Jamshid Shanbehzadeh
65
Cosine Transform Forward Cosine Transform : Inverse Cosine Transform : 65Jamshid Shanbehzadeh
66
The DCT has been used historically in image compression, such as JPEG In computer imaging we often represent the basis matrices as images, called basis images, where we use various gray values to represent the different values in the basis matrix The basis images are separable Cosine Transform 66Jamshid Shanbehzadeh
67
Cosine transform basis functions, N=16. 67Jamshid Shanbehzadeh
68
Cosine transform basis images, N=4. 68Jamshid Shanbehzadeh
69
Cosine Transform 69Jamshid Shanbehzadeh
70
64 تصویر پایه جهت محاسبه تبدیل کسینوسی گسسته: تصویر پایه مربوط به مولفه (3و3) تصویر پایه مربوط به مولفه (7و7) تصویر پایه مربوط به مولفه (5و5) تصویر پایه مربوط به مولفه (7و3) تصاویر پایه تبدیل کسینوسی گسسته 70 Jamshid Shanbehzadeh
71
تصویر اصلی تبدیل کسینوسی 71Jamshid Shanbehzadeh
72
بازسازی تصویر توسط قسمتی از ضرایب DCT الف) بازسازی شده توسط ضرائبی واقع در یک مثلث به ضلع 128 پیکسل (0/125 ضرایب) با 1/1071=MSE ب) بازسازی شده توسط 64 سطر اول و 64 ستون اول (0/4375 ضرایب) با 1/4708=MSE پ) بازسازی شده توسط 32 سطر اول و 32 ستون اول (0/2343 ضرایب) با 1/2087=MSE الف)ب) پ) 72Jamshid Shanbehzadeh
73
بازسازی تصویر توسط قسمتی از ضرایب DCT Leads to Ringing effect Good reconstruction results in applying Cosine Transform in compression 73Jamshid Shanbehzadeh
74
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 74Jamshid Shanbehzadeh
75
Sine Transform 75Jamshid Shanbehzadeh
76
Sine transform basis functions, N=15. 76Jamshid Shanbehzadeh
77
Two-dimensional Sine transform 77Jamshid Shanbehzadeh
78
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 78Jamshid Shanbehzadeh
79
Hartley Transform 79Jamshid Shanbehzadeh
80
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 80Jamshid Shanbehzadeh
81
Hadamard Transform The Hadamard Transform is based on the Hadamard matrix, which is a square array of plus and minus 1s whose rows and columns are orthogonal. 81Jamshid Shanbehzadeh
82
Hadamard Transform A normalized N X N Hadamard matrix satisfies the relation : HH T = 1 82Jamshid Shanbehzadeh
83
Hadamard Transform 83Jamshid Shanbehzadeh
84
Hadamard Transform 84Jamshid Shanbehzadeh
85
Hadamard Transform Basis Function, N=16. حجم تغییرات بسیار بالاست. 85Jamshid Shanbehzadeh
86
Hadamard Transform Basis Images, N=16. 86Jamshid Shanbehzadeh
87
87Jamshid Shanbehzadeh
88
88Jamshid Shanbehzadeh
89
Hadamard Transform 89Jamshid Shanbehzadeh
90
Contents Introduction Applications of Image Transforms Types of Image Transforms 2-D Transform –Basis Image (m 1,m 2 ) –Reverse 2-D Transform Basis Inverse Transform Image (m 1,m 2 ) –2-D Unitary Transform Separable Transform Forward Transform Reverse Transform Unitary Matrix(Transform) Orthogonal Matrix(Transform) Separable Transform –Forward Separable Transform –Inverse Separable Transform Fourier Transform –Forward Fourier Transform –Inverse Fourier Transform –Fourier Transform(Separable) –Fourier Transform Basis Functions –Fourier Transform Properties –Fourier Transform Phase Information –Translation Property –Rotation Property Cosine transform –Basis functions –Basis Images –Cosine symmetry Sine Transform –Basis functions –2-D sine transform Hartley Transform Hadamard Transform –Basis Functions –Basis Images Principle Components Analysis 90Jamshid Shanbehzadeh
91
Principal Component Analysis K= تعداد تصاوير =ابعاد تصاويرmXn تعداد پيکسلهای تصوير N = mXn ابعاد تصوير تبدیلی است مخصوص به یک نوع تصویر خاص مثلا تصویر چهره 91Jamshid Shanbehzadeh
92
92Jamshid Shanbehzadeh
93
ماتریس حاصله ماتریس کواریانس بوده و متقارن می باشد. اگر دو سطر از ماتریس اصلی ناهمبسته باشند، آنگاه عنصر نظیر صفر خواهد شد. 93Jamshid Shanbehzadeh
94
94Jamshid Shanbehzadeh
95
Main Images 95Jamshid Shanbehzadeh
96
Basis Images 96Jamshid Shanbehzadeh
97
Images Generated From 10 Basis Images 97Jamshid Shanbehzadeh
98
Images Generated From 15 Basis Images 98Jamshid Shanbehzadeh
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.